#Write a program to find square of any number in JavaScript

[whizsolutions]

Top Android App Development SDKs, Libraries, and Frameworks to Use in 2021

Google came up with Android in 2007 for the first time and since then it has made our life easier. Many industries like the mobile phone industry have bloomed cause of this and continue doing so. Android has just not got into phones but also, tv, watches, and whatnot.

An Android app development company will always think about how to make an app that incorporates all the customer’s expectations. In addition, there is also the responsibility to see if the app runs smoothly without any lag. So, to help the android developers with these complexities, there are some SDKs, frameworks and libraries that is the most effective till now.

If you are an android developer yourself or are searching for one to aid your business, this list of tools will give you some idea.

Ionic

With this framework, you can develop apps that will be compatible with iOS, Windows and Android. This ionic framework is available free of cost and also gives high mobile performance. It has so many features that its library amazes the developers. Moreover, to code in this framework, you only need to have a basic knowledge of HTML, CSS or JavaScript.

In addition to ionic, you can also use many UI elements while building the application and also use the necessary tab bars to create responsive applications. This tool is a boon to users and developers as it is user friendly and effortless o learn.

Dagger

Dagger is the best dependency injection tool as it is a compile-time framework and thus creates simple java source code. So, what is dependency injection? It is useful to easily inject small components into the larger model. It is better than other dependency injection libraries cause it works in compile-time which allows it to analyze and estimate dependencies.

Moreover, the other dependency injection libraries are limited & depend on XML to function, thus Dagger turns to be the best. Google maintains this tool now, but originally Square created the earlier version of Dagger and this is its adaptation. This is not a user-friendly tool, in fact, a bit difficult to learn, but your learning time will be worth its performance. So, if you are not a pro, hire Android app developer to get your expected work done.

Retrofit

This is a safe and essential HTTP client for working in Java and Android. This one is easy and great for beginners as you can run popular serialization libraries through this. You can define your program’s REST API here as similar to an interface. You need to use annotations to handle API requests body, headers, query parameters, etc.

As we mentioned already, it is simple and user-friendly, and even you can make asynchronous and synchronous API calls here. Also, if you are finding problems in fetching data with HttpsUrlConnection due to its small data capacity, try this Retrofit. This one works very well with RxJava as it already possesses an Rx module of its own.

Glide

It is one of the best libraries with a great memory as it can find out, decode and play videos. When you are finding it difficult to handle images in Android for the image handling API can be strenuous. Use this as the API can crash your app when you go for altering your images.

So, if your goal is to load images or to manage any media, this app is the best in 2021. Any Android app development services company will prefer using this one because of its speed. Moreover, it is an open-source library and while Google recommends it, Bumptech manages it.

Xamarin

You will be able to create a native and modern application in this framework for operating systems like iOS, Android and Windows. Xamarin is very popular among developers as this makes their work easy, and you must try this one. One programming language is enough to work perfectly with Xamarin as it works jointly with a C# code that is supported by any platform.

Also, if you like sharing and testing your code and also want to check its performance through visual studio, look no further. This Xamarin framework can solve all your miseries and it possesses a backend API component too.

Kotlin and Android SDK

This is also one from the bag of developer’s favourites, but semicolons at the end of lines are essential here. If you don’t like to write numerous Java code lines, switch to this for this requires much less boilerplate code. The language of Kotlin and Android SDK is understandable and easy to learn.

So, people prefer this while doing large projects, but Kotlin is only used for developing android applications. Big companies like Uber, Google and Pinterest prefer using it for its attractive features. Its features also include compatibility with Java, less coding, being easily adaptable and no runtime overhead.

Flutter

It is quite new in the field of cross-platform SDKs as it just came in 2017. It works a bit differently than others, But the language it uses is not that new, that is Dart. Dart is there in the programming world for a long time and is a bit matured. This language has many features like higher-order functions, null safety and extension methods.

The language is modern and typesafe and also enable you to create a UI for your users. With Flutter, you can be sure of lag less run in any platform as they are also working to have a position in the desktop platform.

React Native

It is one of the best cross-platform SDKs for building apps for phones. This one is an open-source JavaScript library and helps the developers to build outstanding apps in less than usual time. Facebook developed React Native and it works perfectly fine on any device. Also, it uses JSX, because of which a huge number of JavaScript developers swear by it. These are some of the best SDKs, libraries and frameworks of 2021, though there are more. Visit our website if you need more help on app development, from one of the best Android application development solutions.

[sitesbay]

getElementById

getElementById

In this code we learn about document.getElementById() method in JavaScript.

[t-baba]

Best Free Android Apps and How to Create Them Yourself

You don't need an original app idea to be successful on Google Play. If you spend a few minutes looking at all the app categories there, you'll notice how similar most of the popular apps in any category are. This is especially true in the case of free apps and games.

What I'm trying to say is that you don't always have to spend days brainstorming trying to come up with that perfectly unique app idea. You can make a profit even if you simply create a clone of a popular app, add a few ads to it, and publish it on Google Play. For instance, consider Flappy Bird, an extremely popular game released in 2013. It has innumerable clones today, many of which have millions of downloads. Take a look at Flappy Nyan or Flapping Online if you don't believe me.

In this article, I'll help you find the best free Android apps and games available on Google Play and introduce you to several handcrafted templates you can use to quickly create clones of those apps yourself.

CodeCanyon is a Marketplace for App Templates and Builders

CodeCanyon is an online marketplace that has hundreds of additional professional Android app templates and builder tools. Some of these are incredibly feature-rich and well-designed. You can sometimes save days, even months, of effort by using one of them.

1. WhatsApp Messenger

Messaging apps tend to be very popular on Google Play. With over five billion downloads, WhatsApp Messenger is the poster child for such apps. It supports secure instant messaging, voice and video calls, and allows you to instantly share pictures, videos, and other media with your friends.

To create a WhatsApp Messenger clone, you can use any of the following premium templates:

DreamsChat

DreamsChat is a template for creating real-time chat applications that also support both video and voice calls.

This template has several additional features to match WhatsApp Messenger, such as stickers, group chats, phone number verification, and file sharing.

Plax

Plax is another beautiful template for creating messenger apps for Android. It's fully integrated with Google Firebase, so you don't need to have your own server as a back-end for your apps.

Using this template you can create apps that can do almost everything WhatsApp Messenger can, and more. For instance, it offers text and image stories, in-app notifications, and a dark mode too.

It comes fully integrated with AdMob ads, so you'll have no trouble monetizing your app.

2. Instagram

Instagram is a photo sharing app with over a billion downloads on Google Play. It allows you to upload photos and videos, apply filters to them, and add tags to them so others can find them. Because it's also a social network, it allows you to have followers, write comments, and share stories.

CodeCanyon has some awesome templates you can use to create an Instagram clone in just a few hours. Here are two of them:

My Social Network

My Social Network is a very popular template for creating not only your own social network, but also an Android app for it, which closely resembles Instagram in both looks and functionality.

You'll need PHP and MySQL installed on your server to run the social network, and Android Studio to build the app.

The apps you build with this template will have support for push notifications, AdMob ads, in-app purchases, and even instant messaging.

Pikky

Pikky is another template that lets you quickly create a fully functional Instagram-like app. It uses a Parse server hosted on back4app as its backend, so you don't have to spend any time setting up your own web server or database.

As you may expect, this templates comes with a custom camera implementation that lets you capture square photos and videos, which can be enhanced by applying one of the many filters available.

Pikky supports AdMob interstitial ads.

3. YouTube

Google's YouTube app, with over 5 billion downloads, is among the top 5 most downloaded apps on Google Play. It allows you to view, upload, and edit videos. You can also leave comments on videos, and subscribe to channels to be notified when they have new videos.

Maintaining a video hosting platform is hard because you'll need massive amounts of bandwidth to serve a growing user base. But if you're up for a challenge, you can create your own YouTube clone using the following template:

PlayTube Android

PlayTube Android is a template for creating video sharing apps. It uses PlayTube, a PHP script for creating video sharing social networks, as its back-end.

This template allows your users to upload their own videos, and even import videos from sites like YouTube, Dailymotion, and Vimeo. There's also support for playlists, user channels, and search.

PlayTube Android has AdMob's banner, interstitial, and rewarded ads built into it.

4. Knife Hit

Knife Hit is a free game with over 50 million downloads on Google Play. It has a very simple, yet addictive gameplay: throw knives at rotating logs to split them, while avoiding the spikes on the logs. You get extra points by hitting the apples on the logs.

This game has a few successful clones already, but there's always place for more. If you want to create one, you can use the following premium template:

Knife

Knife is a template that lets you create clones of knife throwing games in a matter of minutes. It comes with several ready-to-use knife skins and game modes. But you can also easily add your own custom game levels.

You won't need any programming skills to replace game assets such as graphics and sounds. But if you're familiar with JavaScript, you can easily add custom functionality and animations too.

The template has well-placed AdMob banner and interstitial ads.

5. Twitter

The Twitter app for Android is a social networking app that has well over 500 million users. It allows you to have a native Android experience while using the Twitter platform. With it, you can effortlessly use your phone to add thoughts, photos, videos, and links to your timeline.

If you're interested in creating a Twitter app clone, here's a template you need to take a look at:

WoWonder Timeline

WoWonder Timeline for Android is a full-fledged template for creating social networking apps. It uses WoWonder, a very popular PHP script for building social networks, as its backend.

This Material Design-compliant template is highly customizable and has everything you'd need to build a successful social networking app. For instance, it supports user profiles, posts, comments, mentions, hashtags, and photo albums.

You can easily use AdMob to monetize the apps you build with this template.

6. Tubi

Tubi is a free, streaming app you can use to watch movies and TV series. It has over 50 million downloads and a growing collection of free shows and movies. It is very similar to Netflix, but you don't need a subscription to be able to watch the content available here.

You can easily create an app similar to Tubi by using the following template:

Flix App

Flix App is a template for creating apps that can stream videos. It allows you to have separate pages for movies, TV series, and live TV channels. It also supports lists for seasons, episodes, and related videos.

The apps you build with this template can play videos in a variety of video formats, such as MP4, M3U, and MOV. They'll also be able to display subtitles so long as you have the required VTT or SRT files.

Flix App has two monetization options built into it: AdMob ads and Facebook Network Audience ads.

7. Google Play Music

Google Play Music is a great choice if you're looking for a radio app. It offers several interesting radio stations, based on genres, artists, and mood. It has over five billion downloads, and dozens of clones, on Google Play.

You can use the following CodeCanyon template to create a clone of Google Play Music:

Android Music Player

Android Music Player is a template for creating music player apps. You can use it create apps that can not only play songs from your server, but also songs that are on users' devices.

This template comes with back-end server code so you can have an admin console for your app. Using the console, you can easily manage your music files. For instance, you can categorize them based on genres and artists, or group them into playlists and albums.

Furthermore, the app template has Firebase Analytics, OneSignal push notifications, and AdMob ads built into it.

Conclusion

According to Statista, over 95% of the apps available on Google Play can be downloaded for free. Paid apps are only a tiny minority there. This is because a free app with ads and in-app purchases, if it can get enough downloads, can rake in significantly higher profits than its paid counterparts.

In this article, you saw how easy it is to create clones of popular, free apps using templates available on CodeCanyon. Go ahead, pick a template, style it to match your preferences, enable AdMob ads, and publish your next app on Google Play.

If you want to learn how to download and customize a CodeCanyon template, check out my other post.

Android

How to Get Started With an Android App Template

Ashraff Hathibelagal

Premium Android App Templates from CodeCanyon

The default templates offered by Android Studio are very basic and provide minimal, generic functionality. CodeCanyon is an online marketplace that has hundreds of additional templates, which are way more feature rich and domain-specific too. 

If you have trouble deciding which template on CodeCanyon is right for you, these articles should help: 

App Templates

20 Best Android App Templates of 2020

Franc Lucas

Android SDK

10 Best Android Game Templates

Ashraff Hathibelagal

App Templates

15 Best eCommerce Android App Templates

Daniel Strongin

Material Design

Best Material Design Android App Templates

Nona Blackman

by Ashraff Hathibelagal via Envato Tuts+ Code https://ift.tt/2UOY690

[shivlab]

Why React Native Development!

Cross-stage improvement has gotten an extraordinary option in contrast to completely Native portable application advancement. Following the Native versatile advancement approach, you make separate applications for Android and iOS. The cross-stage advancement enables you to cut costs and spare time by utilizing a similar code across the two stages. The React Native system is a rising portable arrangement and is viewed as the eventual fate of cross-stage versatile application advancement. 

In this article, we spread the advantages of utilizing React Native for cross-stage versatile improvement and reveal to you when it is anything but a decent decision. 

React Native is an open-source structure that enables you to manufacture a versatile application with the main JavaScript. It was presented by Jordan Walke, a Facebook programming engineer, as another innovation for more straightforward improvement and better client experience. The primary unmistakable of this structure is that React Native Applications work simply like Native applications. They don't contrast from applications based on Java, Objective-C or Quick and they utilize the equivalent UI building hinders as Native iOS or Android applications. In any case, with React Native, fabricating a portable application is a lot quicker and more affordable. 

The measurements show the reasons why an ever-increasing number of engineers decide to React Native for the cross-stage improvement 

Need to know more innovations for your versatile application? 

We should take a gander at the points of interest and a few difficulties of utilizing React Native to build up a portable application.

Advantages of React Native

Network matters 

React Native is an open-source stage. That implies all documentation identified with this innovation is open for everybody and is accessible for nothing to everybody in React Native people group. There's an incredible bit of leeway to utilizing a network-driven innovation. For instance, on the off chance that you face any issue identified with React Native advancement, you can find support from network specialists or discover data on the web. 

For organizations and associations that are thinking about structure a versatile application, React Native is an ideal arrangement that can slice the time and cost down the middle. Besides, if an organization as of now has a web application written in React, a lot of this code can be reused for building a versatile application. 

Streamlined UI 

React Native is about the portable UI. If we contrast this structure with AngularJS or MeteorJS, we find that it looks more like a JavaScript library than a system. 

It's imperative to make a grouping of activities when fabricating a versatile application and React Native makes an actualizing request simply great. Also, UIs planned in React Native are increasingly responsive, decline load time, and give a smoother feel. 

Outsider module support 

The React Native Structure is as yet progressing, so it may come up short on certain segments in the center system. To fill this hole, React Native gives two kinds of outsider modules: Native modules and JavaScript modules. 

For instance, on the off chance that you have to include Google Maps or Google Schedules to your fundamental application usefulness, React Native enables you to connect any module with a Native or outsider module. Among the most noticeable outsider modules for React Native is React Native Selectme, React Native Vector Symbols, React Native Switch Motion, React Native Talented Spinner, React Native Modalbox, and React Native Cabinet. 

Particular design 

Particular writing computer programs is a product plan method that isolates the usefulness of the program into a few autonomous and exchangeable squares called modules. There are two or three favorable circumstances to this procedure: it gives adaptability inside the improvement group as designers can dive into one another's ventures if necessary and it makes it amazingly simple to produce refreshes. React Native natural secluded design helps engineers significantly by giving the capacity to overhaul and refresh applications rapidly. It's additionally conceivable to reuse modules that work both with web and portable APIs. 

Live and hot reloading 

Live and hot reloading are not similar highlights to React Native Application improvement. Allows first to characterize the distinction between these two alternatives. 

Live reloading peruses and orders a document wherein changes were made by an engineer and afterward gives another record to the test system, which consequently reloads the application from the beginning stage. 

Hot reloading depends on Hot Module Substitution (HMR) and was presented after the first reloading alternative. It includes a similar succession of activities, yet when you press Ctrl + S to spare changes, an HMR intermediator embeds the refreshed documents into the necessary spot while the application is running. Perhaps the greatest bit of leeway of hot reload is the capacity to make changes in the source code so they can be seen without the need to recompile the application. On the off chance that a product engineer has two open windows (one with the code and another with the application screen), for example, they can see the outcome on the application screen following applying changes in code. We should take a gander at how this choice functions in the video beneath. 

Definitive coding style 

Definitive programming portrays what the program must do instead of how to do it, which is something contrary to basic programming. A decisive coding style makes React Native code incredibly adaptable and justifiable for engineers. It's additionally gainful when an engineer needs to bounce into another extend and absorb rapidly. The revelatory style rearranges coding ideal models and the coding procedure, bringing about code that is simpler to peruse for both the framework and engineers. A designer can just take a gander at the code and comprehend it due to the splendid UI. 

Instant arrangements 

The React Native structure has a noteworthy rundown of instant arrangements and libraries that extraordinarily encourage portable improvement. Among the enormous number of React Native libraries, we show probably the most valuable ones beneath. 

For fruitful kind checking, there are devices, for example, PropTypes and Stream, while ESLint is an ideal instrument for linting. Axios, React Native firebase, and Apollo Customer are utilized for setting up the systems administration work process to React Native tasks. For a state board, Revival is one of the most as often as possible utilized React Native libraries. 

Moreover, React Native is perfect with JavaScript libraries.

[lakshyadeshwal1-blog]

10 REASONS TO USE NODE.JS

1. Node.js is a Ubiquitous Runtime

JavaScript is a decent spot to begin for new designers, and remains the dialect of decision for some prepared engineers. What’s more, however the dialect has been around for quite a while and put in a couple of years grieving during the 90s and mid 2000s, as intrigue extended, JavaScript has developed and added highlights thought about basic to present day programming dialects. What’s more, similarly as with any great dialect, it is basic and open enough for anybody to begin with, yet ground-breaking enough that even old hands find new things to learn and outfit for their work.

You may as of now be utilizing a rich customer structure, for example, (Angular, Ember, Backbone) and a RESTful server-side API that vans JSON forward and backward. Regardless of whether you’re not utilizing one of those systems, you’ve composed your own in jQuery or Vanilla JavaScript. So in the event that you’re not utilizing Node.js on the server, you’re continually deciphering. You’re deciphering two things: 1) the rationale in your mind from JavaScript to your server-side structure, and 2) the HTTP information from JSON to your server-side articles.

JavaScript as a language is developing in prominence, and has moved outside of the internet browser. It isn’t leaving at any point in the near future.

2. It’s Fast

Node.js is a JavaScript runtime that utilizes the V8 motor created by Google for use in Chrome. V8 accumulates and executes JavaScript at lightning speeds fundamentally because of the way that V8 incorporates JavaScript into local machine code.

Notwithstanding extremely quick JavaScript execution, the genuine enchantment behind Node.js is the occasion circle. The occasion circle is a solitary string that plays out all I/O tasks no concurrently. Generally, I/O tasks either run synchronously (blocking) or no concurrently by bringing forth parallel strings to play out the work. This old methodology devours a great deal of memory and is famously hard to program. Conversely, when a Node application needs to play out an I/O activity, it sends an offbeat errand to the occasion circle, (what is an occasion circle?) alongside a callback capacity, and after that keeps on executing whatever remains of its program. At the point when the a sync activity finishes, the occasion circle comes back to the undertaking to execute its callback.

At the end of the day, perusing and writing to arrange associations, perusing/writing to the file system, and perusing/writing to the database– all exceptionally basic errands in web apps– execute, quick in Node. Utilizing the offbeat example drives you to having an application with elite. Hub enables you to fabricate quick, versatile system applications fit for taking care of an immense number of synchronous associations with high throughput.

3. Modules and Community

npm is the Node.js bundle director and it is magnificent. It does, obviously, take after bundle directors from different biological systems, yet npm is quick, powerful, and steady. It works admirably at determining and introducing venture conditions, yet in addition clouds a lot of complexities. It keeps bundles disengaged from different ventures, dodging form clashes. Be that as it may, it additionally handles worldwide introduces of shell directions and stage subordinate doubles. I can’t recollect a period with npm where I’ve needed to ask myself, “For what reason are those modules clashing?

Where is that module introduced? For what reason is it grabbing this variant and not excessively one?”

snort is the revered undertaking sprinter, yet new children on the square swallow, informal breakfast, and broccoli center around fabricates that change your records, and exploit JavaScript’s solid document streams capacities.

4. JSON in your Database

So you’ve chosen to utilize JavaScript on the server, and you’re glad for your choice that dodges all that making an interpretation of from customer information to server information, yet enduring that information to the database requires considerably more interpretations!

There’s uplifting news. On the off chance that you’re utilizing an article database like Mongo, you can stretch out JavaScript to the industriousness layer too.

Utilizing Node.js enables you to utilize a similar dialect on the customer, on the server, and in the database. You can keep your information in its local JSON arrange from program to plate.

5. Constant, Made Easy with Websockets

On the off chance that Node.js exceeds expectations at numerous simultaneous associations, it bodes well that it exceeds expectations at multi-client, ongoing web applications like talk and diversions. Hub’s occasion circle deals with the multi-client necessity. The ongoing force comes through utilization of the web socket convention. Websockets are basically two-way correspondences channels between the customer and server. So the server can push information to the customer simply as the customer can. Websockets keep running over TCP, evading the overhead of HTTP.

Socket.io is a standout amongst the most prominent web socket libraries being used, and makes shared web applications dead basic.

6. Gushing Data

Generally, web systems treat HTTP solicitations and reactions as entire information objects. Indeed, they’re really I/O streams, as you may get on the off chance that you spilled a document from the file system. Since Node.js is truly adept at dealing with I/O, we can exploit and assemble some cool things. For instance, it’s conceivable to transcode sound or video records while they’re transferring, eliminating the general preparing time.

Hub can peruse/compose streams to web sockets similarly just as it can peruse/compose streams to HTTP. For instance, we can pipe stdout from a running procedure on the server to a program over a web socket, and have the page show the yield progressively.

7. One Codebase and You’re Real-time For Free

On the off chance that you’ve made it this far, you may ask yourself, “If Node.js enables me to compose JavaScript on the customer and server, and makes it simple to send information between the customer and server, would i be able to compose a web application that runs a solitary codebase on both customer and server, and naturally synchronizes information between the two?”

The solution to your inquiry would be truly, and the system for that application would be Meteor.

Meteor is a cutting edge web system worked on Node. It runs the equivalent codebase on the both the customer and server. This enables you to compose customer code that spares straightforwardly to a database. At that point, that information is naturally continued to the server. It works the other way as well! Any information changes on the server are naturally sent to the customer. It shows signs of improvement. Any website page showing that information responds consequently and refreshes itself!

8. Corporate Sponsorships and The Linux Foundation

The intrinsic hazard with any open-source venture is surrender by its volunteer maintainers. This isn’t the situation with Node.js. Hub has corporate sponsorship and worker contribution from Joyent, Microsoft, PayPal, Walmart and others, and is a piece of the Linux Foundation, which drives venture administration and guarantees that Node will appreciate a long, dynamic life.

9. Facilitating Options Galore

With fast appropriation, world-class Node.js facilitating is likewise multiplying. Specifically, Platform-as-a-Service (PaaS) suppliers, for example, Xervo and different hosts lessen arrangements to a solitary order.

10. Each Developer Knows (A Little) JavaScript

This current one’s for your supervisor.

Since the web was in its outset, there has been JavaScript. Each web designer has coded a little JavaScript, regardless of whether that JavaScript was hacking a jQuery module or wiring up a fundamental occasion handler. While picking a web stage, why not pick the stage whose dialect is known by each web engineer on the planet?

For more Information Visit :https://www.skyinfotech.in/nodejs-training-in-noida.php

[isearchgoood]

Machine Learning 101 - Whiteboard Friday

Posted by BritneyMuller

Machine learning is only growing in importance for anyone working in the digital world, but it can often feel like an inaccessible subject. It doesn't have to be — and you don't have to miss out on the competitive edge it can give you when it comes to SEO task automation. Put on your technical SEO cap and get ready to take notes, because Britney Muller is walking us through Machine Learning 101 in this week's episode of Whiteboard Friday.

Click on the whiteboard image above to open a high-resolution version in a new tab!

Video Transcription

Hey, Moz fans. Welcome to another edition of Whiteboard Friday. Today I'm talking about all things machine learning, something, as many of you know, I'm super passionate about and love to talk about. So hopefully, this sparks a seed in some of you to explore it a bit further, because it is truly one of the most powerful things to happen in our space in a very long time. 

What is machine learning?

So a brief overview, in a nutshell, machine learning is actually a subset of AI, and some would argue we still haven't really reached artificial intelligence. But it's just one facet of the overall AI. 

Traditional programming

The best way to think about it is in comparison to traditional programming. So traditional programming, you input data and a program into a computer and out comes the output, whether that be a web page or calculator you built online, whatever that might be.

Machine learning

With machine learning, what you do is you put in the data and the desired output and put this into a computer, and you get a program, otherwise known as a machine learning model. So it's a bit flipped, and it works extremely well. There are two primary types of machine learning:

You have supervised, which is where you're basically feeding a model labeled training data, 

And then unsupervised, which is where you're feeding a program data and letting it create clusters or associations between data points. 

The supervised is a bit more common. You'll see things like classification, linear regression, and image recognition. Things like that are all very common. If you think about machine learning in terms of, okay, there's all of this data that you're putting into the model, data is the biggest part of machine learning. A lot of people would argue that if machine learning was a vehicle, data would be the fuel.

It's a really important part to understand, because unless you have the right types of data to feed a model, you're not going to get the desired outcome that you would like. 

A machine learning model example

So let's look at an example. If you wanted to build a machine learning model that predicts housing prices, you might have all of this information.

You might have the current price, square foot of these homes, land, the number of bathrooms, the number of bedrooms, you name it. It goes on and on. These are also known as features. So what a model is going to try to do, when you put in all of this data, it's going to try to understand associations between this information and come up with a model that best predicts home prices in the future.

The most basic of these machine learning models is linear regression. So if you think about inputting the data where maybe you just put in the price and the square foot, and you can kind of see the data like this. 

You see that as the square foot goes up, so does the price. A model over time, in looking at this data, is going to start to find the smoothest line through the data to have the most accurate predictions in the future.

What you don't want it to do is to fit every single data point and have a line that looks like that — that's also known as overfitting — because it doesn't play nice for new data points. You don't want a model to get so calculated to your dataset that it doesn't predict accurately in the future.

A way to look at that is by the loss function. That's maybe getting a bit deeper in this, but that's how you would measure how the line is being fit. Let's see. 

What are the machine learning possibilities in SEO?

So what are some of the possibilities in SEO? How can we leverage machine learning in the SEO space?

Automate meta descriptions

So there are couple ways that people are already doing this. You can automate meta descriptions by looking at the page content and using a machine model to summarize the text. So this literally summarizes the content for you and pares it down to a meta description length. Pretty incredible. 

Automate titles

You could similarly do this for titles, although I don't suggest you do this for primary pages. This isn't going to be perfect. But if you have a huge, huge website, with hundreds of thousands of pages, it gets you halfway there. It's really interesting to start playing around in that space with these large websites.

Automate image alt text

You can also automate alt text for images. We see these models getting really good at understanding what's in an image. 

Automate 301 redirects

301 redirects, Paul Shapiro has an incredible write-up and basically process for that already. 

Automate content creation

Content creation, and if that scares some of you or if you doubt that these models can currently create content that is decent, I challenge you to go check out Talk to Transformer.

It is a pared-back version of OpenAI, which was founded by Elon Musk. It's pretty incredible and a little scary as to how good the content is just from that pared back model. So that is for sure possible in the future and even today. 

Automate product/page suggestions

In addition to product and page suggestions.

So this is just going to get better. Imagine us providing content and UX specifically for the unique users that come to our site, highly personalized content, highly personalized experiences. Really exciting stuff moving forward. 

Resources

I've got some resources I highly suggest you check out.

Google Codelabs is one of my favorites, just because it walks you through the steps. So if you go to Google Codelabs, filter by TensorFlow or machine learning, you can see the possible examples there. Colab notebooks or Jupyter notebooks are where you'll likely be doing any of the machine learning that you want to do on your own.

Kaggle.com is the number one resource for data science competitions. So you get to really see what are the examples, how are people using machine learning today. You'll see things like TSA has put up over $1 million for a data science team to come up with a model that predicts potential threats from security footage.

This stuff gets really interesting really fast. It's also so important to have diversity and inclusion in this space to avoid really dangerous models in the future. So it's something to definitely think about. 

TensorFlow is a great resource. It's what Google put out, and it's what a lot of their machine learning models is built off of. They've got a really great JavaScript platform that you can play around with. 

Andrew Ng has an incredible machine learning course. I highly suggest you check that out. 

Then Algorithmia is sort of a one-stop shop for models. So if you don't care to dip your toes into machine learning and you just want say a summarizer model or a particular type of model, you could potentially find one there and do a plug-and-play of sorts.



So that's pretty interesting and fun to explore. The last thing is a machine learning model is only as good as the data. I can't express that enough. So a lot of machine learning and data scientists, it's all data cleaning and parsing, and that's the bulk of the work in this field.

It's important to be aware of that. So that's it for Machine Learning 101. Thank you so much for joining me, and I hope to see you all again soon. Thanks.

Video transcription by Speechpad.com

If you enjoyed this episode of Whiteboard Friday, you'll be delighted by all the cutting-edge SEO knowledge you'll get from our newly released MozCon 2019 video bundle. Catch more useful technical tips in Britney's talk, plus 26 additional future-focused topics from our top-notch speakers:

Grab the sessions now!

We suggest scheduling a good old-fashioned knowledge share with your colleagues to educate the whole team — after all, who didn't love movie day in school? ;-)

Sign up for The Moz Top 10, a semimonthly mailer updating you on the top ten hottest pieces of SEO news, tips, and rad links uncovered by the Moz team. Think of it as your exclusive digest of stuff you don't have time to hunt down but want to read!

via Blogger https://ift.tt/2WiKQZt #blogger #bloggingtips #bloggerlife #bloggersgetsocial #ontheblog #writersofinstagram #writingprompt #instapoetry #writerscommunity #writersofig #writersblock #writerlife #writtenword #instawriters #spilledink #wordgasm #creativewriting #poetsofinstagram #blackoutpoetry #poetsofig

[theinjectlikes2]

Machine Learning 101 - Whiteboard Friday

Posted by BritneyMuller

Machine learning is only growing in importance for anyone working in the digital world, but it can often feel like an inaccessible subject. It doesn't have to be — and you don't have to miss out on the competitive edge it can give you when it comes to SEO task automation. Put on your technical SEO cap and get ready to take notes, because Britney Muller is walking us through Machine Learning 101 in this week's episode of Whiteboard Friday.

Click on the whiteboard image above to open a high-resolution version in a new tab!

Video Transcription

Hey, Moz fans. Welcome to another edition of Whiteboard Friday. Today I'm talking about all things machine learning, something, as many of you know, I'm super passionate about and love to talk about. So hopefully, this sparks a seed in some of you to explore it a bit further, because it is truly one of the most powerful things to happen in our space in a very long time. 

What is machine learning?

So a brief overview, in a nutshell, machine learning is actually a subset of AI, and some would argue we still haven't really reached artificial intelligence. But it's just one facet of the overall AI. 

Traditional programming

The best way to think about it is in comparison to traditional programming. So traditional programming, you input data and a program into a computer and out comes the output, whether that be a web page or calculator you built online, whatever that might be.

Machine learning

With machine learning, what you do is you put in the data and the desired output and put this into a computer, and you get a program, otherwise known as a machine learning model. So it's a bit flipped, and it works extremely well. There are two primary types of machine learning:

You have supervised, which is where you're basically feeding a model labeled training data, 

And then unsupervised, which is where you're feeding a program data and letting it create clusters or associations between data points. 

The supervised is a bit more common. You'll see things like classification, linear regression, and image recognition. Things like that are all very common. If you think about machine learning in terms of, okay, there's all of this data that you're putting into the model, data is the biggest part of machine learning. A lot of people would argue that if machine learning was a vehicle, data would be the fuel.

It's a really important part to understand, because unless you have the right types of data to feed a model, you're not going to get the desired outcome that you would like. 

A machine learning model example

So let's look at an example. If you wanted to build a machine learning model that predicts housing prices, you might have all of this information.

You might have the current price, square foot of these homes, land, the number of bathrooms, the number of bedrooms, you name it. It goes on and on. These are also known as features. So what a model is going to try to do, when you put in all of this data, it's going to try to understand associations between this information and come up with a model that best predicts home prices in the future.

The most basic of these machine learning models is linear regression. So if you think about inputting the data where maybe you just put in the price and the square foot, and you can kind of see the data like this. 

You see that as the square foot goes up, so does the price. A model over time, in looking at this data, is going to start to find the smoothest line through the data to have the most accurate predictions in the future.

What you don't want it to do is to fit every single data point and have a line that looks like that — that's also known as overfitting — because it doesn't play nice for new data points. You don't want a model to get so calculated to your dataset that it doesn't predict accurately in the future.

A way to look at that is by the loss function. That's maybe getting a bit deeper in this, but that's how you would measure how the line is being fit. Let's see. 

What are the machine learning possibilities in SEO?

So what are some of the possibilities in SEO? How can we leverage machine learning in the SEO space?

Automate meta descriptions

So there are couple ways that people are already doing this. You can automate meta descriptions by looking at the page content and using a machine model to summarize the text. So this literally summarizes the content for you and pares it down to a meta description length. Pretty incredible. 

Automate titles

You could similarly do this for titles, although I don't suggest you do this for primary pages. This isn't going to be perfect. But if you have a huge, huge website, with hundreds of thousands of pages, it gets you halfway there. It's really interesting to start playing around in that space with these large websites.

Automate image alt text

You can also automate alt text for images. We see these models getting really good at understanding what's in an image. 

Automate 301 redirects

301 redirects, Paul Shapiro has an incredible write-up and basically process for that already. 

Automate content creation

Content creation, and if that scares some of you or if you doubt that these models can currently create content that is decent, I challenge you to go check out Talk to Transformer.

It is a pared-back version of OpenAI, which was founded by Elon Musk. It's pretty incredible and a little scary as to how good the content is just from that pared back model. So that is for sure possible in the future and even today. 

Automate product/page suggestions

In addition to product and page suggestions.

So this is just going to get better. Imagine us providing content and UX specifically for the unique users that come to our site, highly personalized content, highly personalized experiences. Really exciting stuff moving forward. 

Resources

I've got some resources I highly suggest you check out.

Google Codelabs is one of my favorites, just because it walks you through the steps. So if you go to Google Codelabs, filter by TensorFlow or machine learning, you can see the possible examples there. Colab notebooks or Jupyter notebooks are where you'll likely be doing any of the machine learning that you want to do on your own.

Kaggle.com is the number one resource for data science competitions. So you get to really see what are the examples, how are people using machine learning today. You'll see things like TSA has put up over $1 million for a data science team to come up with a model that predicts potential threats from security footage.

This stuff gets really interesting really fast. It's also so important to have diversity and inclusion in this space to avoid really dangerous models in the future. So it's something to definitely think about. 

TensorFlow is a great resource. It's what Google put out, and it's what a lot of their machine learning models is built off of. They've got a really great JavaScript platform that you can play around with. 

Andrew Ng has an incredible machine learning course. I highly suggest you check that out. 

Then Algorithmia is sort of a one-stop shop for models. So if you don't care to dip your toes into machine learning and you just want say a summarizer model or a particular type of model, you could potentially find one there and do a plug-and-play of sorts.



So that's pretty interesting and fun to explore. The last thing is a machine learning model is only as good as the data. I can't express that enough. So a lot of machine learning and data scientists, it's all data cleaning and parsing, and that's the bulk of the work in this field.

It's important to be aware of that. So that's it for Machine Learning 101. Thank you so much for joining me, and I hope to see you all again soon. Thanks.

Video transcription by Speechpad.com

If you enjoyed this episode of Whiteboard Friday, you'll be delighted by all the cutting-edge SEO knowledge you'll get from our newly released MozCon 2019 video bundle. Catch more useful technical tips in Britney's talk, plus 26 additional future-focused topics from our top-notch speakers:

Grab the sessions now!

We suggest scheduling a good old-fashioned knowledge share with your colleagues to educate the whole team — after all, who didn't love movie day in school? ;-)

Sign up for The Moz Top 10, a semimonthly mailer updating you on the top ten hottest pieces of SEO news, tips, and rad links uncovered by the Moz team. Think of it as your exclusive digest of stuff you don't have time to hunt down but want to read!

from The Moz Blog https://ift.tt/31N2yFs via IFTTT

[whitelabelseoreseller]

Machine Learning 101 - Whiteboard Friday

Posted by BritneyMuller

Machine learning is only growing in importance for anyone working in the digital world, but it can often feel like an inaccessible subject. It doesn't have to be — and you don't have to miss out on the competitive edge it can give you when it comes to SEO task automation. Put on your technical SEO cap and get ready to take notes, because Britney Muller is walking us through Machine Learning 101 in this week's episode of Whiteboard Friday.

Click on the whiteboard image above to open a high-resolution version in a new tab!

Video Transcription

Hey, Moz fans. Welcome to another edition of Whiteboard Friday. Today I'm talking about all things machine learning, something, as many of you know, I'm super passionate about and love to talk about. So hopefully, this sparks a seed in some of you to explore it a bit further, because it is truly one of the most powerful things to happen in our space in a very long time. 

What is machine learning?

So a brief overview, in a nutshell, machine learning is actually a subset of AI, and some would argue we still haven't really reached artificial intelligence. But it's just one facet of the overall AI. 

Traditional programming

The best way to think about it is in comparison to traditional programming. So traditional programming, you input data and a program into a computer and out comes the output, whether that be a web page or calculator you built online, whatever that might be.

Machine learning

With machine learning, what you do is you put in the data and the desired output and put this into a computer, and you get a program, otherwise known as a machine learning model. So it's a bit flipped, and it works extremely well. There are two primary types of machine learning:

You have supervised, which is where you're basically feeding a model labeled training data, 

And then unsupervised, which is where you're feeding a program data and letting it create clusters or associations between data points. 

The supervised is a bit more common. You'll see things like classification, linear regression, and image recognition. Things like that are all very common. If you think about machine learning in terms of, okay, there's all of this data that you're putting into the model, data is the biggest part of machine learning. A lot of people would argue that if machine learning was a vehicle, data would be the fuel.

It's a really important part to understand, because unless you have the right types of data to feed a model, you're not going to get the desired outcome that you would like. 

A machine learning model example

So let's look at an example. If you wanted to build a machine learning model that predicts housing prices, you might have all of this information.

You might have the current price, square foot of these homes, land, the number of bathrooms, the number of bedrooms, you name it. It goes on and on. These are also known as features. So what a model is going to try to do, when you put in all of this data, it's going to try to understand associations between this information and come up with a model that best predicts home prices in the future.

The most basic of these machine learning models is linear regression. So if you think about inputting the data where maybe you just put in the price and the square foot, and you can kind of see the data like this. 

You see that as the square foot goes up, so does the price. A model over time, in looking at this data, is going to start to find the smoothest line through the data to have the most accurate predictions in the future.

What you don't want it to do is to fit every single data point and have a line that looks like that — that's also known as overfitting — because it doesn't play nice for new data points. You don't want a model to get so calculated to your dataset that it doesn't predict accurately in the future.

A way to look at that is by the loss function. That's maybe getting a bit deeper in this, but that's how you would measure how the line is being fit. Let's see. 

What are the machine learning possibilities in SEO?

So what are some of the possibilities in SEO? How can we leverage machine learning in the SEO space?

Automate meta descriptions

So there are couple ways that people are already doing this. You can automate meta descriptions by looking at the page content and using a machine model to summarize the text. So this literally summarizes the content for you and pares it down to a meta description length. Pretty incredible. 

Automate titles

You could similarly do this for titles, although I don't suggest you do this for primary pages. This isn't going to be perfect. But if you have a huge, huge website, with hundreds of thousands of pages, it gets you halfway there. It's really interesting to start playing around in that space with these large websites.

Automate image alt text

You can also automate alt text for images. We see these models getting really good at understanding what's in an image. 

Automate 301 redirects

301 redirects, Paul Shapiro has an incredible write-up and basically process for that already. 

Automate content creation

Content creation, and if that scares some of you or if you doubt that these models can currently create content that is decent, I challenge you to go check out Talk to Transformer.

It is a pared-back version of OpenAI, which was founded by Elon Musk. It's pretty incredible and a little scary as to how good the content is just from that pared back model. So that is for sure possible in the future and even today. 

Automate product/page suggestions

In addition to product and page suggestions.

So this is just going to get better. Imagine us providing content and UX specifically for the unique users that come to our site, highly personalized content, highly personalized experiences. Really exciting stuff moving forward. 

Resources

I've got some resources I highly suggest you check out.

Google Codelabs is one of my favorites, just because it walks you through the steps. So if you go to Google Codelabs, filter by TensorFlow or machine learning, you can see the possible examples there. Colab notebooks or Jupyter notebooks are where you'll likely be doing any of the machine learning that you want to do on your own.

Kaggle.com is the number one resource for data science competitions. So you get to really see what are the examples, how are people using machine learning today. You'll see things like TSA has put up over $1 million for a data science team to come up with a model that predicts potential threats from security footage.

This stuff gets really interesting really fast. It's also so important to have diversity and inclusion in this space to avoid really dangerous models in the future. So it's something to definitely think about. 

TensorFlow is a great resource. It's what Google put out, and it's what a lot of their machine learning models is built off of. They've got a really great JavaScript platform that you can play around with. 

Andrew Ng has an incredible machine learning course. I highly suggest you check that out. 

Then Algorithmia is sort of a one-stop shop for models. So if you don't care to dip your toes into machine learning and you just want say a summarizer model or a particular type of model, you could potentially find one there and do a plug-and-play of sorts.



So that's pretty interesting and fun to explore. The last thing is a machine learning model is only as good as the data. I can't express that enough. So a lot of machine learning and data scientists, it's all data cleaning and parsing, and that's the bulk of the work in this field.

It's important to be aware of that. So that's it for Machine Learning 101. Thank you so much for joining me, and I hope to see you all again soon. Thanks.

Video transcription by Speechpad.com

If you enjoyed this episode of Whiteboard Friday, you'll be delighted by all the cutting-edge SEO knowledge you'll get from our newly released MozCon 2019 video bundle. Catch more useful technical tips in Britney's talk, plus 26 additional future-focused topics from our top-notch speakers:

Grab the sessions now!

We suggest scheduling a good old-fashioned knowledge share with your colleagues to educate the whole team — after all, who didn't love movie day in school? ;-)

Sign up for The Moz Top 10, a semimonthly mailer updating you on the top ten hottest pieces of SEO news, tips, and rad links uncovered by the Moz team. Think of it as your exclusive digest of stuff you don't have time to hunt down but want to read!

from The Moz Blog http://tracking.feedpress.it/link/9375/12923731

[bfxenon]

Machine Learning 101 - Whiteboard Friday

Posted by BritneyMuller

Machine learning is only growing in importance for anyone working in the digital world, but it can often feel like an inaccessible subject. It doesn't have to be — and you don't have to miss out on the competitive edge it can give you when it comes to SEO task automation. Put on your technical SEO cap and get ready to take notes, because Britney Muller is walking us through Machine Learning 101 in this week's episode of Whiteboard Friday.

Click on the whiteboard image above to open a high-resolution version in a new tab!

Video Transcription

Hey, Moz fans. Welcome to another edition of Whiteboard Friday. Today I'm talking about all things machine learning, something, as many of you know, I'm super passionate about and love to talk about. So hopefully, this sparks a seed in some of you to explore it a bit further, because it is truly one of the most powerful things to happen in our space in a very long time. 

What is machine learning?

So a brief overview, in a nutshell, machine learning is actually a subset of AI, and some would argue we still haven't really reached artificial intelligence. But it's just one facet of the overall AI. 

Traditional programming

The best way to think about it is in comparison to traditional programming. So traditional programming, you input data and a program into a computer and out comes the output, whether that be a web page or calculator you built online, whatever that might be.

Machine learning

With machine learning, what you do is you put in the data and the desired output and put this into a computer, and you get a program, otherwise known as a machine learning model. So it's a bit flipped, and it works extremely well. There are two primary types of machine learning:

You have supervised, which is where you're basically feeding a model labeled training data, 

And then unsupervised, which is where you're feeding a program data and letting it create clusters or associations between data points. 

The supervised is a bit more common. You'll see things like classification, linear regression, and image recognition. Things like that are all very common. If you think about machine learning in terms of, okay, there's all of this data that you're putting into the model, data is the biggest part of machine learning. A lot of people would argue that if machine learning was a vehicle, data would be the fuel.

It's a really important part to understand, because unless you have the right types of data to feed a model, you're not going to get the desired outcome that you would like. 

A machine learning model example

So let's look at an example. If you wanted to build a machine learning model that predicts housing prices, you might have all of this information.

You might have the current price, square foot of these homes, land, the number of bathrooms, the number of bedrooms, you name it. It goes on and on. These are also known as features. So what a model is going to try to do, when you put in all of this data, it's going to try to understand associations between this information and come up with a model that best predicts home prices in the future.

The most basic of these machine learning models is linear regression. So if you think about inputting the data where maybe you just put in the price and the square foot, and you can kind of see the data like this. 

You see that as the square foot goes up, so does the price. A model over time, in looking at this data, is going to start to find the smoothest line through the data to have the most accurate predictions in the future.

What you don't want it to do is to fit every single data point and have a line that looks like that — that's also known as overfitting — because it doesn't play nice for new data points. You don't want a model to get so calculated to your dataset that it doesn't predict accurately in the future.

A way to look at that is by the loss function. That's maybe getting a bit deeper in this, but that's how you would measure how the line is being fit. Let's see. 

What are the machine learning possibilities in SEO?

So what are some of the possibilities in SEO? How can we leverage machine learning in the SEO space?

Automate meta descriptions

So there are couple ways that people are already doing this. You can automate meta descriptions by looking at the page content and using a machine model to summarize the text. So this literally summarizes the content for you and pares it down to a meta description length. Pretty incredible. 

Automate titles

You could similarly do this for titles, although I don't suggest you do this for primary pages. This isn't going to be perfect. But if you have a huge, huge website, with hundreds of thousands of pages, it gets you halfway there. It's really interesting to start playing around in that space with these large websites.

Automate image alt text

You can also automate alt text for images. We see these models getting really good at understanding what's in an image. 

Automate 301 redirects

301 redirects, Paul Shapiro has an incredible write-up and basically process for that already. 

Automate content creation

Content creation, and if that scares some of you or if you doubt that these models can currently create content that is decent, I challenge you to go check out Talk to Transformer.

It is a pared-back version of OpenAI, which was founded by Elon Musk. It's pretty incredible and a little scary as to how good the content is just from that pared back model. So that is for sure possible in the future and even today. 

Automate product/page suggestions

In addition to product and page suggestions.

So this is just going to get better. Imagine us providing content and UX specifically for the unique users that come to our site, highly personalized content, highly personalized experiences. Really exciting stuff moving forward. 

Resources

I've got some resources I highly suggest you check out.

Google Codelabs is one of my favorites, just because it walks you through the steps. So if you go to Google Codelabs, filter by TensorFlow or machine learning, you can see the possible examples there. Colab notebooks or Jupyter notebooks are where you'll likely be doing any of the machine learning that you want to do on your own.

Kaggle.com is the number one resource for data science competitions. So you get to really see what are the examples, how are people using machine learning today. You'll see things like TSA has put up over $1 million for a data science team to come up with a model that predicts potential threats from security footage.

This stuff gets really interesting really fast. It's also so important to have diversity and inclusion in this space to avoid really dangerous models in the future. So it's something to definitely think about. 

TensorFlow is a great resource. It's what Google put out, and it's what a lot of their machine learning models is built off of. They've got a really great JavaScript platform that you can play around with. 

Andrew Ng has an incredible machine learning course. I highly suggest you check that out. 

Then Algorithmia is sort of a one-stop shop for models. So if you don't care to dip your toes into machine learning and you just want say a summarizer model or a particular type of model, you could potentially find one there and do a plug-and-play of sorts.



So that's pretty interesting and fun to explore. The last thing is a machine learning model is only as good as the data. I can't express that enough. So a lot of machine learning and data scientists, it's all data cleaning and parsing, and that's the bulk of the work in this field.

It's important to be aware of that. So that's it for Machine Learning 101. Thank you so much for joining me, and I hope to see you all again soon. Thanks.

Video transcription by Speechpad.com

If you enjoyed this episode of Whiteboard Friday, you'll be delighted by all the cutting-edge SEO knowledge you'll get from our newly released MozCon 2019 video bundle. Catch more useful technical tips in Britney's talk, plus 26 additional future-focused topics from our top-notch speakers:

Grab the sessions now!

We suggest scheduling a good old-fashioned knowledge share with your colleagues to educate the whole team — after all, who didn't love movie day in school? ;-)

Sign up for The Moz Top 10, a semimonthly mailer updating you on the top ten hottest pieces of SEO news, tips, and rad links uncovered by the Moz team. Think of it as your exclusive digest of stuff you don't have time to hunt down but want to read!

[kjt-lawyers]

Machine Learning 101 - Whiteboard Friday

Posted by BritneyMuller

Machine learning is only growing in importance for anyone working in the digital world, but it can often feel like an inaccessible subject. It doesn't have to be — and you don't have to miss out on the competitive edge it can give you when it comes to SEO task automation. Put on your technical SEO cap and get ready to take notes, because Britney Muller is walking us through Machine Learning 101 in this week's episode of Whiteboard Friday.

Click on the whiteboard image above to open a high-resolution version in a new tab!

Video Transcription

Hey, Moz fans. Welcome to another edition of Whiteboard Friday. Today I'm talking about all things machine learning, something, as many of you know, I'm super passionate about and love to talk about. So hopefully, this sparks a seed in some of you to explore it a bit further, because it is truly one of the most powerful things to happen in our space in a very long time. 

What is machine learning?

So a brief overview, in a nutshell, machine learning is actually a subset of AI, and some would argue we still haven't really reached artificial intelligence. But it's just one facet of the overall AI. 

Traditional programming

The best way to think about it is in comparison to traditional programming. So traditional programming, you input data and a program into a computer and out comes the output, whether that be a web page or calculator you built online, whatever that might be.

Machine learning

With machine learning, what you do is you put in the data and the desired output and put this into a computer, and you get a program, otherwise known as a machine learning model. So it's a bit flipped, and it works extremely well. There are two primary types of machine learning:

You have supervised, which is where you're basically feeding a model labeled training data, 

And then unsupervised, which is where you're feeding a program data and letting it create clusters or associations between data points. 

The supervised is a bit more common. You'll see things like classification, linear regression, and image recognition. Things like that are all very common. If you think about machine learning in terms of, okay, there's all of this data that you're putting into the model, data is the biggest part of machine learning. A lot of people would argue that if machine learning was a vehicle, data would be the fuel.

It's a really important part to understand, because unless you have the right types of data to feed a model, you're not going to get the desired outcome that you would like. 

A machine learning model example

So let's look at an example. If you wanted to build a machine learning model that predicts housing prices, you might have all of this information.

You might have the current price, square foot of these homes, land, the number of bathrooms, the number of bedrooms, you name it. It goes on and on. These are also known as features. So what a model is going to try to do, when you put in all of this data, it's going to try to understand associations between this information and come up with a model that best predicts home prices in the future.

The most basic of these machine learning models is linear regression. So if you think about inputting the data where maybe you just put in the price and the square foot, and you can kind of see the data like this. 

You see that as the square foot goes up, so does the price. A model over time, in looking at this data, is going to start to find the smoothest line through the data to have the most accurate predictions in the future.

What you don't want it to do is to fit every single data point and have a line that looks like that — that's also known as overfitting — because it doesn't play nice for new data points. You don't want a model to get so calculated to your dataset that it doesn't predict accurately in the future.

A way to look at that is by the loss function. That's maybe getting a bit deeper in this, but that's how you would measure how the line is being fit. Let's see. 

What are the machine learning possibilities in SEO?

So what are some of the possibilities in SEO? How can we leverage machine learning in the SEO space?

Automate meta descriptions

So there are couple ways that people are already doing this. You can automate meta descriptions by looking at the page content and using a machine model to summarize the text. So this literally summarizes the content for you and pares it down to a meta description length. Pretty incredible. 

Automate titles

You could similarly do this for titles, although I don't suggest you do this for primary pages. This isn't going to be perfect. But if you have a huge, huge website, with hundreds of thousands of pages, it gets you halfway there. It's really interesting to start playing around in that space with these large websites.

Automate image alt text

You can also automate alt text for images. We see these models getting really good at understanding what's in an image. 

Automate 301 redirects

301 redirects, Paul Shapiro has an incredible write-up and basically process for that already. 

Automate content creation

Content creation, and if that scares some of you or if you doubt that these models can currently create content that is decent, I challenge you to go check out Talk to Transformer.

It is a pared-back version of OpenAI, which was founded by Elon Musk. It's pretty incredible and a little scary as to how good the content is just from that pared back model. So that is for sure possible in the future and even today. 

Automate product/page suggestions

In addition to product and page suggestions.

So this is just going to get better. Imagine us providing content and UX specifically for the unique users that come to our site, highly personalized content, highly personalized experiences. Really exciting stuff moving forward. 

Resources

I've got some resources I highly suggest you check out.

Google Codelabs is one of my favorites, just because it walks you through the steps. So if you go to Google Codelabs, filter by TensorFlow or machine learning, you can see the possible examples there. Colab notebooks or Jupyter notebooks are where you'll likely be doing any of the machine learning that you want to do on your own.

Kaggle.com is the number one resource for data science competitions. So you get to really see what are the examples, how are people using machine learning today. You'll see things like TSA has put up over $1 million for a data science team to come up with a model that predicts potential threats from security footage.

This stuff gets really interesting really fast. It's also so important to have diversity and inclusion in this space to avoid really dangerous models in the future. So it's something to definitely think about. 

TensorFlow is a great resource. It's what Google put out, and it's what a lot of their machine learning models is built off of. They've got a really great JavaScript platform that you can play around with. 

Andrew Ng has an incredible machine learning course. I highly suggest you check that out. 

Then Algorithmia is sort of a one-stop shop for models. So if you don't care to dip your toes into machine learning and you just want say a summarizer model or a particular type of model, you could potentially find one there and do a plug-and-play of sorts.



So that's pretty interesting and fun to explore. The last thing is a machine learning model is only as good as the data. I can't express that enough. So a lot of machine learning and data scientists, it's all data cleaning and parsing, and that's the bulk of the work in this field.

It's important to be aware of that. So that's it for Machine Learning 101. Thank you so much for joining me, and I hope to see you all again soon. Thanks.

Video transcription by Speechpad.com

If you enjoyed this episode of Whiteboard Friday, you'll be delighted by all the cutting-edge SEO knowledge you'll get from our newly released MozCon 2019 video bundle. Catch more useful technical tips in Britney's talk, plus 26 additional future-focused topics from our top-notch speakers:

Grab the sessions now!

We suggest scheduling a good old-fashioned knowledge share with your colleagues to educate the whole team — after all, who didn't love movie day in school? ;-)

Sign up for The Moz Top 10, a semimonthly mailer updating you on the top ten hottest pieces of SEO news, tips, and rad links uncovered by the Moz team. Think of it as your exclusive digest of stuff you don't have time to hunt down but want to read!

[noithatotoaz]

Machine Learning 101 - Whiteboard Friday

Posted by BritneyMuller

Machine learning is only growing in importance for anyone working in the digital world, but it can often feel like an inaccessible subject. It doesn't have to be — and you don't have to miss out on the competitive edge it can give you when it comes to SEO task automation. Put on your technical SEO cap and get ready to take notes, because Britney Muller is walking us through Machine Learning 101 in this week's episode of Whiteboard Friday.

Click on the whiteboard image above to open a high-resolution version in a new tab!

Video Transcription

Hey, Moz fans. Welcome to another edition of Whiteboard Friday. Today I'm talking about all things machine learning, something, as many of you know, I'm super passionate about and love to talk about. So hopefully, this sparks a seed in some of you to explore it a bit further, because it is truly one of the most powerful things to happen in our space in a very long time. 

What is machine learning?

So a brief overview, in a nutshell, machine learning is actually a subset of AI, and some would argue we still haven't really reached artificial intelligence. But it's just one facet of the overall AI. 

Traditional programming

The best way to think about it is in comparison to traditional programming. So traditional programming, you input data and a program into a computer and out comes the output, whether that be a web page or calculator you built online, whatever that might be.

Machine learning

With machine learning, what you do is you put in the data and the desired output and put this into a computer, and you get a program, otherwise known as a machine learning model. So it's a bit flipped, and it works extremely well. There are two primary types of machine learning:

You have supervised, which is where you're basically feeding a model labeled training data, 

And then unsupervised, which is where you're feeding a program data and letting it create clusters or associations between data points. 

The supervised is a bit more common. You'll see things like classification, linear regression, and image recognition. Things like that are all very common. If you think about machine learning in terms of, okay, there's all of this data that you're putting into the model, data is the biggest part of machine learning. A lot of people would argue that if machine learning was a vehicle, data would be the fuel.

It's a really important part to understand, because unless you have the right types of data to feed a model, you're not going to get the desired outcome that you would like. 

A machine learning model example

So let's look at an example. If you wanted to build a machine learning model that predicts housing prices, you might have all of this information.

You might have the current price, square foot of these homes, land, the number of bathrooms, the number of bedrooms, you name it. It goes on and on. These are also known as features. So what a model is going to try to do, when you put in all of this data, it's going to try to understand associations between this information and come up with a model that best predicts home prices in the future.

The most basic of these machine learning models is linear regression. So if you think about inputting the data where maybe you just put in the price and the square foot, and you can kind of see the data like this. 

You see that as the square foot goes up, so does the price. A model over time, in looking at this data, is going to start to find the smoothest line through the data to have the most accurate predictions in the future.

What you don't want it to do is to fit every single data point and have a line that looks like that — that's also known as overfitting — because it doesn't play nice for new data points. You don't want a model to get so calculated to your dataset that it doesn't predict accurately in the future.

A way to look at that is by the loss function. That's maybe getting a bit deeper in this, but that's how you would measure how the line is being fit. Let's see. 

What are the machine learning possibilities in SEO?

So what are some of the possibilities in SEO? How can we leverage machine learning in the SEO space?

Automate meta descriptions

So there are couple ways that people are already doing this. You can automate meta descriptions by looking at the page content and using a machine model to summarize the text. So this literally summarizes the content for you and pares it down to a meta description length. Pretty incredible. 

Automate titles

You could similarly do this for titles, although I don't suggest you do this for primary pages. This isn't going to be perfect. But if you have a huge, huge website, with hundreds of thousands of pages, it gets you halfway there. It's really interesting to start playing around in that space with these large websites.

Automate image alt text

You can also automate alt text for images. We see these models getting really good at understanding what's in an image. 

Automate 301 redirects

301 redirects, Paul Shapiro has an incredible write-up and basically process for that already. 

Automate content creation

Content creation, and if that scares some of you or if you doubt that these models can currently create content that is decent, I challenge you to go check out Talk to Transformer.

It is a pared-back version of OpenAI, which was founded by Elon Musk. It's pretty incredible and a little scary as to how good the content is just from that pared back model. So that is for sure possible in the future and even today. 

Automate product/page suggestions

In addition to product and page suggestions.

So this is just going to get better. Imagine us providing content and UX specifically for the unique users that come to our site, highly personalized content, highly personalized experiences. Really exciting stuff moving forward. 

Resources

I've got some resources I highly suggest you check out.

Google Codelabs is one of my favorites, just because it walks you through the steps. So if you go to Google Codelabs, filter by TensorFlow or machine learning, you can see the possible examples there. Colab notebooks or Jupyter notebooks are where you'll likely be doing any of the machine learning that you want to do on your own.

Kaggle.com is the number one resource for data science competitions. So you get to really see what are the examples, how are people using machine learning today. You'll see things like TSA has put up over $1 million for a data science team to come up with a model that predicts potential threats from security footage.

This stuff gets really interesting really fast. It's also so important to have diversity and inclusion in this space to avoid really dangerous models in the future. So it's something to definitely think about. 

TensorFlow is a great resource. It's what Google put out, and it's what a lot of their machine learning models is built off of. They've got a really great JavaScript platform that you can play around with. 

Andrew Ng has an incredible machine learning course. I highly suggest you check that out. 

Then Algorithmia is sort of a one-stop shop for models. So if you don't care to dip your toes into machine learning and you just want say a summarizer model or a particular type of model, you could potentially find one there and do a plug-and-play of sorts.



So that's pretty interesting and fun to explore. The last thing is a machine learning model is only as good as the data. I can't express that enough. So a lot of machine learning and data scientists, it's all data cleaning and parsing, and that's the bulk of the work in this field.

It's important to be aware of that. So that's it for Machine Learning 101. Thank you so much for joining me, and I hope to see you all again soon. Thanks.

Video transcription by Speechpad.com

If you enjoyed this episode of Whiteboard Friday, you'll be delighted by all the cutting-edge SEO knowledge you'll get from our newly released MozCon 2019 video bundle. Catch more useful technical tips in Britney's talk, plus 26 additional future-focused topics from our top-notch speakers:

Grab the sessions now!

We suggest scheduling a good old-fashioned knowledge share with your colleagues to educate the whole team — after all, who didn't love movie day in school? ;-)

Sign up for The Moz Top 10, a semimonthly mailer updating you on the top ten hottest pieces of SEO news, tips, and rad links uncovered by the Moz team. Think of it as your exclusive digest of stuff you don't have time to hunt down but want to read!

[thanhtuandoan89]

Machine Learning 101 - Whiteboard Friday

Posted by BritneyMuller

Machine learning is only growing in importance for anyone working in the digital world, but it can often feel like an inaccessible subject. It doesn't have to be — and you don't have to miss out on the competitive edge it can give you when it comes to SEO task automation. Put on your technical SEO cap and get ready to take notes, because Britney Muller is walking us through Machine Learning 101 in this week's episode of Whiteboard Friday.

Click on the whiteboard image above to open a high-resolution version in a new tab!

Video Transcription

Hey, Moz fans. Welcome to another edition of Whiteboard Friday. Today I'm talking about all things machine learning, something, as many of you know, I'm super passionate about and love to talk about. So hopefully, this sparks a seed in some of you to explore it a bit further, because it is truly one of the most powerful things to happen in our space in a very long time. 

What is machine learning?

So a brief overview, in a nutshell, machine learning is actually a subset of AI, and some would argue we still haven't really reached artificial intelligence. But it's just one facet of the overall AI. 

Traditional programming

The best way to think about it is in comparison to traditional programming. So traditional programming, you input data and a program into a computer and out comes the output, whether that be a web page or calculator you built online, whatever that might be.

Machine learning

With machine learning, what you do is you put in the data and the desired output and put this into a computer, and you get a program, otherwise known as a machine learning model. So it's a bit flipped, and it works extremely well. There are two primary types of machine learning:

You have supervised, which is where you're basically feeding a model labeled training data, 

And then unsupervised, which is where you're feeding a program data and letting it create clusters or associations between data points. 

The supervised is a bit more common. You'll see things like classification, linear regression, and image recognition. Things like that are all very common. If you think about machine learning in terms of, okay, there's all of this data that you're putting into the model, data is the biggest part of machine learning. A lot of people would argue that if machine learning was a vehicle, data would be the fuel.

It's a really important part to understand, because unless you have the right types of data to feed a model, you're not going to get the desired outcome that you would like. 

A machine learning model example

So let's look at an example. If you wanted to build a machine learning model that predicts housing prices, you might have all of this information.

You might have the current price, square foot of these homes, land, the number of bathrooms, the number of bedrooms, you name it. It goes on and on. These are also known as features. So what a model is going to try to do, when you put in all of this data, it's going to try to understand associations between this information and come up with a model that best predicts home prices in the future.

The most basic of these machine learning models is linear regression. So if you think about inputting the data where maybe you just put in the price and the square foot, and you can kind of see the data like this. 

You see that as the square foot goes up, so does the price. A model over time, in looking at this data, is going to start to find the smoothest line through the data to have the most accurate predictions in the future.

What you don't want it to do is to fit every single data point and have a line that looks like that — that's also known as overfitting — because it doesn't play nice for new data points. You don't want a model to get so calculated to your dataset that it doesn't predict accurately in the future.

A way to look at that is by the loss function. That's maybe getting a bit deeper in this, but that's how you would measure how the line is being fit. Let's see. 

What are the machine learning possibilities in SEO?

So what are some of the possibilities in SEO? How can we leverage machine learning in the SEO space?

Automate meta descriptions

So there are couple ways that people are already doing this. You can automate meta descriptions by looking at the page content and using a machine model to summarize the text. So this literally summarizes the content for you and pares it down to a meta description length. Pretty incredible. 

Automate titles

You could similarly do this for titles, although I don't suggest you do this for primary pages. This isn't going to be perfect. But if you have a huge, huge website, with hundreds of thousands of pages, it gets you halfway there. It's really interesting to start playing around in that space with these large websites.

Automate image alt text

You can also automate alt text for images. We see these models getting really good at understanding what's in an image. 

Automate 301 redirects

301 redirects, Paul Shapiro has an incredible write-up and basically process for that already. 

Automate content creation

Content creation, and if that scares some of you or if you doubt that these models can currently create content that is decent, I challenge you to go check out Talk to Transformer.

It is a pared-back version of OpenAI, which was founded by Elon Musk. It's pretty incredible and a little scary as to how good the content is just from that pared back model. So that is for sure possible in the future and even today. 

Automate product/page suggestions

In addition to product and page suggestions.

So this is just going to get better. Imagine us providing content and UX specifically for the unique users that come to our site, highly personalized content, highly personalized experiences. Really exciting stuff moving forward. 

Resources

I've got some resources I highly suggest you check out.

Google Codelabs is one of my favorites, just because it walks you through the steps. So if you go to Google Codelabs, filter by TensorFlow or machine learning, you can see the possible examples there. Colab notebooks or Jupyter notebooks are where you'll likely be doing any of the machine learning that you want to do on your own.

Kaggle.com is the number one resource for data science competitions. So you get to really see what are the examples, how are people using machine learning today. You'll see things like TSA has put up over $1 million for a data science team to come up with a model that predicts potential threats from security footage.

This stuff gets really interesting really fast. It's also so important to have diversity and inclusion in this space to avoid really dangerous models in the future. So it's something to definitely think about. 

TensorFlow is a great resource. It's what Google put out, and it's what a lot of their machine learning models is built off of. They've got a really great JavaScript platform that you can play around with. 

Andrew Ng has an incredible machine learning course. I highly suggest you check that out. 

Then Algorithmia is sort of a one-stop shop for models. So if you don't care to dip your toes into machine learning and you just want say a summarizer model or a particular type of model, you could potentially find one there and do a plug-and-play of sorts.



So that's pretty interesting and fun to explore. The last thing is a machine learning model is only as good as the data. I can't express that enough. So a lot of machine learning and data scientists, it's all data cleaning and parsing, and that's the bulk of the work in this field.

It's important to be aware of that. So that's it for Machine Learning 101. Thank you so much for joining me, and I hope to see you all again soon. Thanks.

Video transcription by Speechpad.com

If you enjoyed this episode of Whiteboard Friday, you'll be delighted by all the cutting-edge SEO knowledge you'll get from our newly released MozCon 2019 video bundle. Catch more useful technical tips in Britney's talk, plus 26 additional future-focused topics from our top-notch speakers:

Grab the sessions now!

We suggest scheduling a good old-fashioned knowledge share with your colleagues to educate the whole team — after all, who didn't love movie day in school? ;-)

Sign up for The Moz Top 10, a semimonthly mailer updating you on the top ten hottest pieces of SEO news, tips, and rad links uncovered by the Moz team. Think of it as your exclusive digest of stuff you don't have time to hunt down but want to read!

[ductrungnguyen87]

Machine Learning 101 - Whiteboard Friday

Posted by BritneyMuller

Machine learning is only growing in importance for anyone working in the digital world, but it can often feel like an inaccessible subject. It doesn't have to be — and you don't have to miss out on the competitive edge it can give you when it comes to SEO task automation. Put on your technical SEO cap and get ready to take notes, because Britney Muller is walking us through Machine Learning 101 in this week's episode of Whiteboard Friday.

Click on the whiteboard image above to open a high-resolution version in a new tab!

Video Transcription

Hey, Moz fans. Welcome to another edition of Whiteboard Friday. Today I'm talking about all things machine learning, something, as many of you know, I'm super passionate about and love to talk about. So hopefully, this sparks a seed in some of you to explore it a bit further, because it is truly one of the most powerful things to happen in our space in a very long time. 

What is machine learning?

So a brief overview, in a nutshell, machine learning is actually a subset of AI, and some would argue we still haven't really reached artificial intelligence. But it's just one facet of the overall AI. 

Traditional programming

The best way to think about it is in comparison to traditional programming. So traditional programming, you input data and a program into a computer and out comes the output, whether that be a web page or calculator you built online, whatever that might be.

Machine learning

With machine learning, what you do is you put in the data and the desired output and put this into a computer, and you get a program, otherwise known as a machine learning model. So it's a bit flipped, and it works extremely well. There are two primary types of machine learning:

You have supervised, which is where you're basically feeding a model labeled training data, 

And then unsupervised, which is where you're feeding a program data and letting it create clusters or associations between data points. 

The supervised is a bit more common. You'll see things like classification, linear regression, and image recognition. Things like that are all very common. If you think about machine learning in terms of, okay, there's all of this data that you're putting into the model, data is the biggest part of machine learning. A lot of people would argue that if machine learning was a vehicle, data would be the fuel.

It's a really important part to understand, because unless you have the right types of data to feed a model, you're not going to get the desired outcome that you would like. 

A machine learning model example

So let's look at an example. If you wanted to build a machine learning model that predicts housing prices, you might have all of this information.

You might have the current price, square foot of these homes, land, the number of bathrooms, the number of bedrooms, you name it. It goes on and on. These are also known as features. So what a model is going to try to do, when you put in all of this data, it's going to try to understand associations between this information and come up with a model that best predicts home prices in the future.

The most basic of these machine learning models is linear regression. So if you think about inputting the data where maybe you just put in the price and the square foot, and you can kind of see the data like this. 

You see that as the square foot goes up, so does the price. A model over time, in looking at this data, is going to start to find the smoothest line through the data to have the most accurate predictions in the future.

What you don't want it to do is to fit every single data point and have a line that looks like that — that's also known as overfitting — because it doesn't play nice for new data points. You don't want a model to get so calculated to your dataset that it doesn't predict accurately in the future.

A way to look at that is by the loss function. That's maybe getting a bit deeper in this, but that's how you would measure how the line is being fit. Let's see. 

What are the machine learning possibilities in SEO?

So what are some of the possibilities in SEO? How can we leverage machine learning in the SEO space?

Automate meta descriptions

So there are couple ways that people are already doing this. You can automate meta descriptions by looking at the page content and using a machine model to summarize the text. So this literally summarizes the content for you and pares it down to a meta description length. Pretty incredible. 

Automate titles

You could similarly do this for titles, although I don't suggest you do this for primary pages. This isn't going to be perfect. But if you have a huge, huge website, with hundreds of thousands of pages, it gets you halfway there. It's really interesting to start playing around in that space with these large websites.

Automate image alt text

You can also automate alt text for images. We see these models getting really good at understanding what's in an image. 

Automate 301 redirects

301 redirects, Paul Shapiro has an incredible write-up and basically process for that already. 

Automate content creation

Content creation, and if that scares some of you or if you doubt that these models can currently create content that is decent, I challenge you to go check out Talk to Transformer.

It is a pared-back version of OpenAI, which was founded by Elon Musk. It's pretty incredible and a little scary as to how good the content is just from that pared back model. So that is for sure possible in the future and even today. 

Automate product/page suggestions

In addition to product and page suggestions.

So this is just going to get better. Imagine us providing content and UX specifically for the unique users that come to our site, highly personalized content, highly personalized experiences. Really exciting stuff moving forward. 

Resources

I've got some resources I highly suggest you check out.

Google Codelabs is one of my favorites, just because it walks you through the steps. So if you go to Google Codelabs, filter by TensorFlow or machine learning, you can see the possible examples there. Colab notebooks or Jupyter notebooks are where you'll likely be doing any of the machine learning that you want to do on your own.

Kaggle.com is the number one resource for data science competitions. So you get to really see what are the examples, how are people using machine learning today. You'll see things like TSA has put up over $1 million for a data science team to come up with a model that predicts potential threats from security footage.

This stuff gets really interesting really fast. It's also so important to have diversity and inclusion in this space to avoid really dangerous models in the future. So it's something to definitely think about. 

TensorFlow is a great resource. It's what Google put out, and it's what a lot of their machine learning models is built off of. They've got a really great JavaScript platform that you can play around with. 

Andrew Ng has an incredible machine learning course. I highly suggest you check that out. 

Then Algorithmia is sort of a one-stop shop for models. So if you don't care to dip your toes into machine learning and you just want say a summarizer model or a particular type of model, you could potentially find one there and do a plug-and-play of sorts.



So that's pretty interesting and fun to explore. The last thing is a machine learning model is only as good as the data. I can't express that enough. So a lot of machine learning and data scientists, it's all data cleaning and parsing, and that's the bulk of the work in this field.

It's important to be aware of that. So that's it for Machine Learning 101. Thank you so much for joining me, and I hope to see you all again soon. Thanks.

Video transcription by Speechpad.com

If you enjoyed this episode of Whiteboard Friday, you'll be delighted by all the cutting-edge SEO knowledge you'll get from our newly released MozCon 2019 video bundle. Catch more useful technical tips in Britney's talk, plus 26 additional future-focused topics from our top-notch speakers:

Grab the sessions now!

We suggest scheduling a good old-fashioned knowledge share with your colleagues to educate the whole team — after all, who didn't love movie day in school? ;-)

Sign up for The Moz Top 10, a semimonthly mailer updating you on the top ten hottest pieces of SEO news, tips, and rad links uncovered by the Moz team. Think of it as your exclusive digest of stuff you don't have time to hunt down but want to read!

[gamebazu]

Machine Learning 101 - Whiteboard Friday

Posted by BritneyMuller

Machine learning is only growing in importance for anyone working in the digital world, but it can often feel like an inaccessible subject. It doesn't have to be — and you don't have to miss out on the competitive edge it can give you when it comes to SEO task automation. Put on your technical SEO cap and get ready to take notes, because Britney Muller is walking us through Machine Learning 101 in this week's episode of Whiteboard Friday.

Click on the whiteboard image above to open a high-resolution version in a new tab!

Video Transcription

Hey, Moz fans. Welcome to another edition of Whiteboard Friday. Today I'm talking about all things machine learning, something, as many of you know, I'm super passionate about and love to talk about. So hopefully, this sparks a seed in some of you to explore it a bit further, because it is truly one of the most powerful things to happen in our space in a very long time. 

What is machine learning?

So a brief overview, in a nutshell, machine learning is actually a subset of AI, and some would argue we still haven't really reached artificial intelligence. But it's just one facet of the overall AI. 

Traditional programming

The best way to think about it is in comparison to traditional programming. So traditional programming, you input data and a program into a computer and out comes the output, whether that be a web page or calculator you built online, whatever that might be.

Machine learning

With machine learning, what you do is you put in the data and the desired output and put this into a computer, and you get a program, otherwise known as a machine learning model. So it's a bit flipped, and it works extremely well. There are two primary types of machine learning:

You have supervised, which is where you're basically feeding a model labeled training data, 

And then unsupervised, which is where you're feeding a program data and letting it create clusters or associations between data points. 

The supervised is a bit more common. You'll see things like classification, linear regression, and image recognition. Things like that are all very common. If you think about machine learning in terms of, okay, there's all of this data that you're putting into the model, data is the biggest part of machine learning. A lot of people would argue that if machine learning was a vehicle, data would be the fuel.

It's a really important part to understand, because unless you have the right types of data to feed a model, you're not going to get the desired outcome that you would like. 

A machine learning model example

So let's look at an example. If you wanted to build a machine learning model that predicts housing prices, you might have all of this information.

You might have the current price, square foot of these homes, land, the number of bathrooms, the number of bedrooms, you name it. It goes on and on. These are also known as features. So what a model is going to try to do, when you put in all of this data, it's going to try to understand associations between this information and come up with a model that best predicts home prices in the future.

The most basic of these machine learning models is linear regression. So if you think about inputting the data where maybe you just put in the price and the square foot, and you can kind of see the data like this. 

You see that as the square foot goes up, so does the price. A model over time, in looking at this data, is going to start to find the smoothest line through the data to have the most accurate predictions in the future.

What you don't want it to do is to fit every single data point and have a line that looks like that — that's also known as overfitting — because it doesn't play nice for new data points. You don't want a model to get so calculated to your dataset that it doesn't predict accurately in the future.

A way to look at that is by the loss function. That's maybe getting a bit deeper in this, but that's how you would measure how the line is being fit. Let's see. 

What are the machine learning possibilities in SEO?

So what are some of the possibilities in SEO? How can we leverage machine learning in the SEO space?

Automate meta descriptions

So there are couple ways that people are already doing this. You can automate meta descriptions by looking at the page content and using a machine model to summarize the text. So this literally summarizes the content for you and pares it down to a meta description length. Pretty incredible. 

Automate titles

You could similarly do this for titles, although I don't suggest you do this for primary pages. This isn't going to be perfect. But if you have a huge, huge website, with hundreds of thousands of pages, it gets you halfway there. It's really interesting to start playing around in that space with these large websites.

Automate image alt text

You can also automate alt text for images. We see these models getting really good at understanding what's in an image. 

Automate 301 redirects

301 redirects, Paul Shapiro has an incredible write-up and basically process for that already. 

Automate content creation

Content creation, and if that scares some of you or if you doubt that these models can currently create content that is decent, I challenge you to go check out Talk to Transformer.

It is a pared-back version of OpenAI, which was founded by Elon Musk. It's pretty incredible and a little scary as to how good the content is just from that pared back model. So that is for sure possible in the future and even today. 

Automate product/page suggestions

In addition to product and page suggestions.

So this is just going to get better. Imagine us providing content and UX specifically for the unique users that come to our site, highly personalized content, highly personalized experiences. Really exciting stuff moving forward. 

Resources

I've got some resources I highly suggest you check out.

Google Codelabs is one of my favorites, just because it walks you through the steps. So if you go to Google Codelabs, filter by TensorFlow or machine learning, you can see the possible examples there. Colab notebooks or Jupyter notebooks are where you'll likely be doing any of the machine learning that you want to do on your own.

Kaggle.com is the number one resource for data science competitions. So you get to really see what are the examples, how are people using machine learning today. You'll see things like TSA has put up over $1 million for a data science team to come up with a model that predicts potential threats from security footage.

This stuff gets really interesting really fast. It's also so important to have diversity and inclusion in this space to avoid really dangerous models in the future. So it's something to definitely think about. 

TensorFlow is a great resource. It's what Google put out, and it's what a lot of their machine learning models is built off of. They've got a really great JavaScript platform that you can play around with. 

Andrew Ng has an incredible machine learning course. I highly suggest you check that out. 

Then Algorithmia is sort of a one-stop shop for models. So if you don't care to dip your toes into machine learning and you just want say a summarizer model or a particular type of model, you could potentially find one there and do a plug-and-play of sorts.



So that's pretty interesting and fun to explore. The last thing is a machine learning model is only as good as the data. I can't express that enough. So a lot of machine learning and data scientists, it's all data cleaning and parsing, and that's the bulk of the work in this field.

It's important to be aware of that. So that's it for Machine Learning 101. Thank you so much for joining me, and I hope to see you all again soon. Thanks.

Video transcription by Speechpad.com

If you enjoyed this episode of Whiteboard Friday, you'll be delighted by all the cutting-edge SEO knowledge you'll get from our newly released MozCon 2019 video bundle. Catch more useful technical tips in Britney's talk, plus 26 additional future-focused topics from our top-notch speakers:

Grab the sessions now!

We suggest scheduling a good old-fashioned knowledge share with your colleagues to educate the whole team — after all, who didn't love movie day in school? ;-)

Sign up for The Moz Top 10, a semimonthly mailer updating you on the top ten hottest pieces of SEO news, tips, and rad links uncovered by the Moz team. Think of it as your exclusive digest of stuff you don't have time to hunt down but want to read!

https://ift.tt/2Pj0DGo

[camerasieunhovn]

Machine Learning 101 - Whiteboard Friday

Posted by BritneyMuller

Machine learning is only growing in importance for anyone working in the digital world, but it can often feel like an inaccessible subject. It doesn't have to be — and you don't have to miss out on the competitive edge it can give you when it comes to SEO task automation. Put on your technical SEO cap and get ready to take notes, because Britney Muller is walking us through Machine Learning 101 in this week's episode of Whiteboard Friday.

Click on the whiteboard image above to open a high-resolution version in a new tab!

Video Transcription

Hey, Moz fans. Welcome to another edition of Whiteboard Friday. Today I'm talking about all things machine learning, something, as many of you know, I'm super passionate about and love to talk about. So hopefully, this sparks a seed in some of you to explore it a bit further, because it is truly one of the most powerful things to happen in our space in a very long time. 

What is machine learning?

So a brief overview, in a nutshell, machine learning is actually a subset of AI, and some would argue we still haven't really reached artificial intelligence. But it's just one facet of the overall AI. 

Traditional programming

The best way to think about it is in comparison to traditional programming. So traditional programming, you input data and a program into a computer and out comes the output, whether that be a web page or calculator you built online, whatever that might be.

Machine learning

With machine learning, what you do is you put in the data and the desired output and put this into a computer, and you get a program, otherwise known as a machine learning model. So it's a bit flipped, and it works extremely well. There are two primary types of machine learning:

You have supervised, which is where you're basically feeding a model labeled training data, 

And then unsupervised, which is where you're feeding a program data and letting it create clusters or associations between data points. 

The supervised is a bit more common. You'll see things like classification, linear regression, and image recognition. Things like that are all very common. If you think about machine learning in terms of, okay, there's all of this data that you're putting into the model, data is the biggest part of machine learning. A lot of people would argue that if machine learning was a vehicle, data would be the fuel.

It's a really important part to understand, because unless you have the right types of data to feed a model, you're not going to get the desired outcome that you would like. 

A machine learning model example

So let's look at an example. If you wanted to build a machine learning model that predicts housing prices, you might have all of this information.

You might have the current price, square foot of these homes, land, the number of bathrooms, the number of bedrooms, you name it. It goes on and on. These are also known as features. So what a model is going to try to do, when you put in all of this data, it's going to try to understand associations between this information and come up with a model that best predicts home prices in the future.

The most basic of these machine learning models is linear regression. So if you think about inputting the data where maybe you just put in the price and the square foot, and you can kind of see the data like this. 

You see that as the square foot goes up, so does the price. A model over time, in looking at this data, is going to start to find the smoothest line through the data to have the most accurate predictions in the future.

What you don't want it to do is to fit every single data point and have a line that looks like that — that's also known as overfitting — because it doesn't play nice for new data points. You don't want a model to get so calculated to your dataset that it doesn't predict accurately in the future.

A way to look at that is by the loss function. That's maybe getting a bit deeper in this, but that's how you would measure how the line is being fit. Let's see. 

What are the machine learning possibilities in SEO?

So what are some of the possibilities in SEO? How can we leverage machine learning in the SEO space?

Automate meta descriptions

So there are couple ways that people are already doing this. You can automate meta descriptions by looking at the page content and using a machine model to summarize the text. So this literally summarizes the content for you and pares it down to a meta description length. Pretty incredible. 

Automate titles

You could similarly do this for titles, although I don't suggest you do this for primary pages. This isn't going to be perfect. But if you have a huge, huge website, with hundreds of thousands of pages, it gets you halfway there. It's really interesting to start playing around in that space with these large websites.

Automate image alt text

You can also automate alt text for images. We see these models getting really good at understanding what's in an image. 

Automate 301 redirects

301 redirects, Paul Shapiro has an incredible write-up and basically process for that already. 

Automate content creation

Content creation, and if that scares some of you or if you doubt that these models can currently create content that is decent, I challenge you to go check out Talk to Transformer.

It is a pared-back version of OpenAI, which was founded by Elon Musk. It's pretty incredible and a little scary as to how good the content is just from that pared back model. So that is for sure possible in the future and even today. 

Automate product/page suggestions

In addition to product and page suggestions.

So this is just going to get better. Imagine us providing content and UX specifically for the unique users that come to our site, highly personalized content, highly personalized experiences. Really exciting stuff moving forward. 

Resources

I've got some resources I highly suggest you check out.

Google Codelabs is one of my favorites, just because it walks you through the steps. So if you go to Google Codelabs, filter by TensorFlow or machine learning, you can see the possible examples there. Colab notebooks or Jupyter notebooks are where you'll likely be doing any of the machine learning that you want to do on your own.

Kaggle.com is the number one resource for data science competitions. So you get to really see what are the examples, how are people using machine learning today. You'll see things like TSA has put up over $1 million for a data science team to come up with a model that predicts potential threats from security footage.

This stuff gets really interesting really fast. It's also so important to have diversity and inclusion in this space to avoid really dangerous models in the future. So it's something to definitely think about. 

TensorFlow is a great resource. It's what Google put out, and it's what a lot of their machine learning models is built off of. They've got a really great JavaScript platform that you can play around with. 

Andrew Ng has an incredible machine learning course. I highly suggest you check that out. 

Then Algorithmia is sort of a one-stop shop for models. So if you don't care to dip your toes into machine learning and you just want say a summarizer model or a particular type of model, you could potentially find one there and do a plug-and-play of sorts.



So that's pretty interesting and fun to explore. The last thing is a machine learning model is only as good as the data. I can't express that enough. So a lot of machine learning and data scientists, it's all data cleaning and parsing, and that's the bulk of the work in this field.

It's important to be aware of that. So that's it for Machine Learning 101. Thank you so much for joining me, and I hope to see you all again soon. Thanks.

Video transcription by Speechpad.com

If you enjoyed this episode of Whiteboard Friday, you'll be delighted by all the cutting-edge SEO knowledge you'll get from our newly released MozCon 2019 video bundle. Catch more useful technical tips in Britney's talk, plus 26 additional future-focused topics from our top-notch speakers:

Grab the sessions now!

We suggest scheduling a good old-fashioned knowledge share with your colleagues to educate the whole team — after all, who didn't love movie day in school? ;-)

Sign up for The Moz Top 10, a semimonthly mailer updating you on the top ten hottest pieces of SEO news, tips, and rad links uncovered by the Moz team. Think of it as your exclusive digest of stuff you don't have time to hunt down but want to read!

[drummcarpentry]

Machine Learning 101 - Whiteboard Friday

Posted by BritneyMuller

Machine learning is only growing in importance for anyone working in the digital world, but it can often feel like an inaccessible subject. It doesn't have to be — and you don't have to miss out on the competitive edge it can give you when it comes to SEO task automation. Put on your technical SEO cap and get ready to take notes, because Britney Muller is walking us through Machine Learning 101 in this week's episode of Whiteboard Friday.

Click on the whiteboard image above to open a high-resolution version in a new tab!

Video Transcription

Hey, Moz fans. Welcome to another edition of Whiteboard Friday. Today I'm talking about all things machine learning, something, as many of you know, I'm super passionate about and love to talk about. So hopefully, this sparks a seed in some of you to explore it a bit further, because it is truly one of the most powerful things to happen in our space in a very long time. 

What is machine learning?

So a brief overview, in a nutshell, machine learning is actually a subset of AI, and some would argue we still haven't really reached artificial intelligence. But it's just one facet of the overall AI. 

Traditional programming

The best way to think about it is in comparison to traditional programming. So traditional programming, you input data and a program into a computer and out comes the output, whether that be a web page or calculator you built online, whatever that might be.

Machine learning

With machine learning, what you do is you put in the data and the desired output and put this into a computer, and you get a program, otherwise known as a machine learning model. So it's a bit flipped, and it works extremely well. There are two primary types of machine learning:

You have supervised, which is where you're basically feeding a model labeled training data, 

And then unsupervised, which is where you're feeding a program data and letting it create clusters or associations between data points. 

The supervised is a bit more common. You'll see things like classification, linear regression, and image recognition. Things like that are all very common. If you think about machine learning in terms of, okay, there's all of this data that you're putting into the model, data is the biggest part of machine learning. A lot of people would argue that if machine learning was a vehicle, data would be the fuel.

It's a really important part to understand, because unless you have the right types of data to feed a model, you're not going to get the desired outcome that you would like. 

A machine learning model example

So let's look at an example. If you wanted to build a machine learning model that predicts housing prices, you might have all of this information.

You might have the current price, square foot of these homes, land, the number of bathrooms, the number of bedrooms, you name it. It goes on and on. These are also known as features. So what a model is going to try to do, when you put in all of this data, it's going to try to understand associations between this information and come up with a model that best predicts home prices in the future.

The most basic of these machine learning models is linear regression. So if you think about inputting the data where maybe you just put in the price and the square foot, and you can kind of see the data like this. 

You see that as the square foot goes up, so does the price. A model over time, in looking at this data, is going to start to find the smoothest line through the data to have the most accurate predictions in the future.

What you don't want it to do is to fit every single data point and have a line that looks like that — that's also known as overfitting — because it doesn't play nice for new data points. You don't want a model to get so calculated to your dataset that it doesn't predict accurately in the future.

A way to look at that is by the loss function. That's maybe getting a bit deeper in this, but that's how you would measure how the line is being fit. Let's see. 

What are the machine learning possibilities in SEO?

So what are some of the possibilities in SEO? How can we leverage machine learning in the SEO space?

Automate meta descriptions

So there are couple ways that people are already doing this. You can automate meta descriptions by looking at the page content and using a machine model to summarize the text. So this literally summarizes the content for you and pares it down to a meta description length. Pretty incredible. 

Automate titles

You could similarly do this for titles, although I don't suggest you do this for primary pages. This isn't going to be perfect. But if you have a huge, huge website, with hundreds of thousands of pages, it gets you halfway there. It's really interesting to start playing around in that space with these large websites.

Automate image alt text

You can also automate alt text for images. We see these models getting really good at understanding what's in an image. 

Automate 301 redirects

301 redirects, Paul Shapiro has an incredible write-up and basically process for that already. 

Automate content creation

Content creation, and if that scares some of you or if you doubt that these models can currently create content that is decent, I challenge you to go check out Talk to Transformer.

It is a pared-back version of OpenAI, which was founded by Elon Musk. It's pretty incredible and a little scary as to how good the content is just from that pared back model. So that is for sure possible in the future and even today. 

Automate product/page suggestions

In addition to product and page suggestions.

So this is just going to get better. Imagine us providing content and UX specifically for the unique users that come to our site, highly personalized content, highly personalized experiences. Really exciting stuff moving forward. 

Resources

I've got some resources I highly suggest you check out.

Google Codelabs is one of my favorites, just because it walks you through the steps. So if you go to Google Codelabs, filter by TensorFlow or machine learning, you can see the possible examples there. Colab notebooks or Jupyter notebooks are where you'll likely be doing any of the machine learning that you want to do on your own.

Kaggle.com is the number one resource for data science competitions. So you get to really see what are the examples, how are people using machine learning today. You'll see things like TSA has put up over $1 million for a data science team to come up with a model that predicts potential threats from security footage.

This stuff gets really interesting really fast. It's also so important to have diversity and inclusion in this space to avoid really dangerous models in the future. So it's something to definitely think about. 

TensorFlow is a great resource. It's what Google put out, and it's what a lot of their machine learning models is built off of. They've got a really great JavaScript platform that you can play around with. 

Andrew Ng has an incredible machine learning course. I highly suggest you check that out. 

Then Algorithmia is sort of a one-stop shop for models. So if you don't care to dip your toes into machine learning and you just want say a summarizer model or a particular type of model, you could potentially find one there and do a plug-and-play of sorts.



So that's pretty interesting and fun to explore. The last thing is a machine learning model is only as good as the data. I can't express that enough. So a lot of machine learning and data scientists, it's all data cleaning and parsing, and that's the bulk of the work in this field.

It's important to be aware of that. So that's it for Machine Learning 101. Thank you so much for joining me, and I hope to see you all again soon. Thanks.

Video transcription by Speechpad.com

If you enjoyed this episode of Whiteboard Friday, you'll be delighted by all the cutting-edge SEO knowledge you'll get from our newly released MozCon 2019 video bundle. Catch more useful technical tips in Britney's talk, plus 26 additional future-focused topics from our top-notch speakers:

Grab the sessions now!

We suggest scheduling a good old-fashioned knowledge share with your colleagues to educate the whole team — after all, who didn't love movie day in school? ;-)

Sign up for The Moz Top 10, a semimonthly mailer updating you on the top ten hottest pieces of SEO news, tips, and rad links uncovered by the Moz team. Think of it as your exclusive digest of stuff you don't have time to hunt down but want to read!