#inconsistent-data

[codewithnazam]

Cleaning Dirty Data in Python: Practical Techniques with Pandas

I. Introduction Hey there! So, let’s talk about a really important step in data analysis: data cleaning. It’s basically like tidying up your room before a big party – you want everything to be neat and organized so you can find what you need, right? Now, when it comes to sorting through a bunch of messy data, you’ll be glad to have a tool like Pandas by your side. It’s like the superhero of…

[bird-everyday]

424+425

[millenniallust4death]

Recall: My original goal was to write R code to back up my Blaze post in case Tumblr disappears.

Here’s what I learned about Tumblr’s rickety API based on a test post with ~800 notes.

What I thought I would see:

A beautiful reblog tree, clearly showing how my post spread across Tumblr — every reblog linked to the one it came from.

Posts with tags would show up as blue dots in my version tree. (Tags interest me because they represent people who more deeply engaged with my post. And I'm nosy. XD)

What Tumblr’s rickety API actually gave me:

Orphaned nodes. Reblogs that definitely happened but aren’t connected to anything.

I expected to get a clean version history: Original Post (Blog A) → Blog B → Blog C

But the API fucker just shrugs and tells you C reblogged A. What the fuck happened to Blog B?

The result is that whole branches of the tree vanish. And that’s in addition to the reblogs that are missing because of the API respecting blog privacy settings (as it definitely should).

Conclusion:

As always, I feel disappointed but not surprised by Tumblr’s inability to fully function. The data doesn’t reflect what I know Tumblr users actually did; it just reflects what the API is willing to share.

I am so tired.

TL;DR: Nope. I cannot fully back up my Blaze post using the rickety API.

A quick rendering of the reblog version tree from the test post — a visual representation of my disappointment.

[demifiendcruithne]

today on digital devil saga 2 is criminally under-researched (aka crui is a nerd about a 20 year old game)

so i finally got the impulse/opportunity to test out something i've been suspecting for a while but have never seen any documentation on: dds2 has some form of STAB* mechanic with elemental spells for playable characters, based on their associated element (the one they resist + the one their unique gun combos use)

*for non-pokemon folk: STAB == same-type attack bonus. so for example, a chimecho (psychic type pokemon) using extrasensory (psychic type move) will deal more damage than a non-psychic-type pokemon with the same special attack against the same target

i have more tests i intend to make when i have spoons (i'm just grinding mantras for the superboss and noticed it while fighting wild narasimha because that hecker disables my berserker ring), but so far i've tested with my three active characters fulfilling the following:

are fighting the exact same narasimha who has no elemental weakness/resistance/etc (savestates were used and run multiple times; the rng is not fixed until a move is actually used, so i could get multiple data points quickly by making a new savestate every time i loaded it to advance the rng)

have 99 magic

are level 99

have no elemental boost or amp skills equipped

have no rings equipped that may affect damage, outside of boosting magic stat to 99 for consistency

use two different dynes; one of their associated element, and one not. the former consistently did more, and with repeated runs i cannot write it off as damage range.

under cut: many images; by virtue of being in endgame, these will have some spoilers for endgame in plain sight. if you're reading this and haven't gotten to the final dungeon, i highly recommend you go do that and come back later. once you've reached the final dungeon, you're safe to check it out.

(but seriously - i can't stop you, but i highly recommend actually. playing this game first. it's great)

if you're reading this, you've either gotten to the final dungeon or don't care about spoilers; in case of the latter, i'll do a quick explanation of our last party member momentarily. i have been a little misleading because they are. very much a BIG spoiler. but first - gale resists force, and does more damage with zandyne than bufudyne; likewise argilla with earth/teradyne/ziodyne.

(bufudyne was a random pick for gale because i had only five for omoikane reasons, but in hindsight it was a good one...)

now - for those unaware, this last party member's name is seraph. they're a fusion of two other party members both of said party members were associated with/resisted ice, and i'm near-certain that they also got the boost to bufu spells (i can check this at some point, i have a backup save right before endgame) the same way gale and argilla do.

but what's seraph's elemental affinities? no weaknesses, no resistances. so what does this mean here?

extra screenshots because the first agidyne was a high roll - literally the highest i've seen on a non-stab unboosted dyne against a narasimha. yet it's still clearly estranged from gale's zandyne and argilla's teradyne.

the other part? seraph's bufudyne does the same as gale's, and the same as argilla's ziodyne and their own agidyne, the same as expected without stab.

i'm not sure what else to say about this thus far besides putting a list of further tests to do eventually / get proper data for:

100s of attempts so it can be statistically analysed better. easy to code the data collection into seraph (my helper program, named after this seraph because. yeah), possibly more difficult to analyse but my partner got me into this game he can deal with the consequences (being asked to do data stuff) :v

including cielo and heat as well, probably after i beat satan and grind them up as a victory lap (in the mantras sense not the internment facility sense)

earlier save reloads for data for serph, sera, and roland - i doubt they'd be any different, but regardless

other elemental spells - both the basic and -la levels, the ma-s, the maxed out mantra ones, and the dances. is the boost approximately consistent?

does it affect combos? i believe it doesn't, same as amp skills and the like (and because i feel like i would have noticed by now), but will get data for both participants initiating

does it affect items? i feel like that's unlikely, but worth trying

is there a 'reverse-stab' for someone's weakness? i don't think so, but i'll get the data for free and if there's a trend i just got to test the ailment damagers for cielo

likewise, is the elemental boost innate like the gun combo elements? almost certainly. but it would be funny if equipping x resist made your x do more damage. and also not really that helpful?

does seraph get a boost on almighty? my first instinct would be yes, purely because of seraph's gun combo doing almighty damage; it's near-useless if you got reincarnate but still worth checking. and if so, does ragnarok behave differently from megido+/last word?

on that note, does seraph resist almighty? i... wouldn't be surprised - status only says no weakness, not no resistance, but almighty resistance tends to be invisible anyway (looking at you, devil survivor 2 that shows everything else) (you're fine just using as an example)

that would be a lot harder to test but if i get a party of seraph / someone with dragon ring / someone else with all maxed stats and just find something with megido+

anything else i might have forgotten

if you hack pyriphlegethon onto heat does it do more damage. i don't think i can test this one but i would. honestly be curious

but yeah. this is pretty cool, maybe i'll check for it in dds1 as well. if anyone else checks any of this out and makes a writeup or anything please ping me i have autistic hyperfocused on jrpg mechanics for over two decades :v

[celestialcass]

For Bot/Human Constructs, Internal Storage Space is vital. Placing the communicator given to you by your Mutual Administrative Assistant in the top spot, close to your heart... it says a certain something special about your *gag* relationship

Bonus Long Haired MB + close ups under the cut

[consumed-star]

Here's a drawing of the b o y I did like a month ago

[fairyhagmother]

Somethings I think a lot about:

1. How Scottish / English ballads often depict acts of violence perpetrated by women, but American ballads often depict acts of violence perpetrated by men (exception being young hunting and probably others but the shift in emphasis does feel quite remarkable to me)

2. American ballads will often edit out the ‘retribution’ part of the Scottish / English ballad they are adapting which is like. sad bc maybe I want to see a man haunted by the ghost of the lover he killed and then be ripped to bits by her. who wouldn’t? but also I love the idea of the perpetration of the song as a form of retribution / haunting.

Much to think about

[appavevo]

okay bear with me here:

in the tng episode "datalore" from the 1st season, data mentions that he can tell some work station belonged to "dr. soong." to which both riker and la forge are like who? and wait, you don't mean NOONIEN soong right?

so from that dialogue, it stands to reason that the fact that noonien soon is the creator of data either wasn't known, or wasn't common knowledge. Since data is super unique and mentioned in several biomechanical textbooks, i feel its unlikely that riker wouldn't know about noonien soong, especially since riker said in read through data's file in the pilot.

ANYWAY, this sort of suggests that it wasn't known that dr noonien soong was the creator of data. which is a funny idea to me

BECAUSE

so in the first season of tng, data says he was found 26 years ago. we know from that episode that noonien soong failed to create a positronic brain, and disappeared. after disappearing, he was successful to varying degrees, with b4/lore/data. so i'm going to estimate that his disappearance happened about 40-50 years before the episode "datalore", and about 20-30 years before data was discovered right?

so imagine: you are a middle age cyberneticist working at star fleet. you remember noonien soong from a long time ago, as a weird but smart man who seemed to not be able to handle failure, but that was a long time ago, no sense thinking about it now. then you hear the news: an artificial being has been discovered, possibly one with sentience. This being has requested to join star fleet, and there are currently trials being conducted to determine whether or not to permit this. You think whoa! i gotta go check this out. so you see this mysterious android, whom no one seems to know the origin of, and you think huh......this guy

sure does look familiar......?

eh, probably just a coincidence.....

[watery-melon-baller]

does anyone know a better way to try and reformat a research question for a survey because I am SUFFERING

[thelastspeecher]

God I fucking hate reviewing SOPs at work

People in STEM are absolutely horrible at writing

[beesorcery]

eyes emoji @ your pacific rim spreadsheets. may I see it 👁👁👁👁👁

:D eheheee

[shinraelectricpowercom]

"ao3 getting rid of the piped silm tags? good" no you don't get it, this is a data-structural CATASTROPHE. some of the character tags have been depiped and some of the relationship tags have been depiped but not all of them, and it's not even consistent across the two. i'll be writing to my senator.

[soupacool]

wondering how many people out there are dissociative as fuck and flying under the radar because of ADHD

[mitsde123]

A Beginner’s Guide to Data Cleaning Techniques

Data is the lifeblood of any modern organization. However, raw data is rarely ready for analysis. Before it can be used for insights, data must be cleaned, refined, and structured—a process known as data cleaning. This blog will explore essential data-cleaning techniques, why they are important, and how beginners can master them.

[some-programming-pearls]

How to handle data inconsistencies across microservices?

Scenario: Imagine a shopping application with multiple microservices handling different aspects such as inventory management, order processing, and payment handling. A customer places an order for a product that is in stock, triggering the deduction of the item from the inventory microservice. However, due to network delays or system failures, the order microservice fails to update the order…

[exopelagic]

fuck Hamilton (bill) all the homies hate Hamilton (bill)