danielwatts99-blog - Tumblr blog

[danielwatts99-blog]

28/05/2021

So today is the day, I submitted everything and it's been a hell of a rollercoaster, I have learned so much throughout this projects. I didn't even know how to write one line of code in python before this project, and now I have a whole recommender system which I will genuinely use. I have learned so much about testing and how important user feedback is, also how to push your system to the limits with edge case tests. I have also improved my English skills as writing a 12,000 word report was extremely difficult and this helped my time management skills and organisation.

Thank you for reading my blog and following me on this journey.

For next week I will -

*Relax for a week :-)

Once again thank you,

Daniel Watts

[danielwatts99-blog]

21/05/2021

So today was my last meeting with Tim, he said just touch up your report and get ready to submit, we also went over the presentation video so I need to create a little powerpoint presentation, and a video just saying my motivation for the project, how I went about the project and a short demonstration of my system.

For next week I will -

*Make video presentation

*Touch up report

*Submit :-)

[danielwatts99-blog]

14/05/2021

So today is the day (sort of) I have completed my report and now have a full draft. In my meeting Tim told me just to get rid of some unnecessary white spacing, refer tests to what specification requirement they link to and show both sides of the report in the conclusion.

For next week I will -

* Remove any unnecessary white spacing

* Link test cases back to the specification

*Show both the success and failures in the conclusion chapter

[danielwatts99-blog]

07/04/2021

We have a week until the official deadline, it was extended until the 28th but I am trying to get it all done by the 14th and from then spend time just touching up minor bits. I completed a lot of work this week, completing the design chapter and the implementation chapter, now it's time to work on the testing and evaluation which again is a chunky part of my report as I need to put my white box testing, black box testing and edge case testing which I am quite proud of.

For next week I will

*Finish up report

*Write testing and evaluation chapter

*Write conclusion chapter

[danielwatts99-blog]

30/04/2021

The end is near. I now have two weeks until the final submission date, my meeting with Tim went through my report and we changed part of my introduction and he said that in my report I need to be objective. Meaning that I need to refrain from talking in first person when talking about the design of the system. This means that the report is more focused on the actual system rather than me as the designer. I have completed until halfway through the design chapter which is quite thick.

For next week I will

*Finish the design chapter

*Draw system architecture graphs to show how my system works

*Add the relevant references to my text

*Begin the implementation chapter

[danielwatts99-blog]

23/04/2021

So now I have only a few weeks until the final submission which is on the 14th of may, so I have to start thinking about the presentation video which is taking place instead of the vivas, as it worth 5% I am going to focus on the report which obviously takes priority, I have created a draft up until the specification meaning I have written up the background research and the introduction, I think my background research is quite strong as I have researched a lot of current recommender systems, how they work and the benefits of them.

For next week I will -

*Work on my specification part of the report

[danielwatts99-blog]

16/04/2021

This week I have made some significant progress on my report, I have gathered the background research I completed in the early phase of the project before the design and incorporated this into my report, where I have looked at previous recommendation systems as well as common python libraries to help.

Next week I will -

Continue work on the report

[danielwatts99-blog]

09/04/2021

The user evaluation has come back, this was really interesting to see because it measures how well my system has performed against my expectations and the specification, the results came back very well, with hardly any dissatisfaction scores and it appears the system is successful in generating recommendations for movies.

For next week I will - 

*Create a document with all of the user surveys so that when the testing and evaluation section of my report is ready the data is easily transferred.

Continue work on Report

[danielwatts99-blog]

02/04/2021

So this week I have finished up my testing for the report, I have decided to do blackbox testing, white box testing and me and Tim discussed the possibility of edge case testing, which essentially pushes my system to the limit and tries to make the system fail. So this has been done for the report which is coming along well, I have been working on this a lot and will continue to do this throughout the Easter holidays.

For next week I will:

*Continue working on report

*Issue the system to users for user evaluation

[danielwatts99-blog]

28/03/21

This week has been a slow week, I have had final assignments due for two modules security and digital venture creation so work on my final project has not been too productive this week although I have still made some developments.

For example I have carried out some white box testing, which means testing the input and output

I have also worked on the draft report which is due soon.

For next week I will -

continue work on the report

continue my testing

[danielwatts99-blog]

21/03/2021

update

This is my last week of term and therefore my last meeting with Tim until we start meeting next term to go over the report, we agreed that the system should not change significantly and stick with what I have, although I would like to have done more, Tim said that what I have done so far is good and I accomplished essentially the aim of the project which is to recommend movies. I have started work on the report.

For next week I will

Continue work on my report

Continue work on my testing

[danielwatts99-blog]

14/03/2021

Update

So now I only have two months until the submission of my report, so I am looking to get started on that very soon, I have done some more data analysis and improved the accuracy score by stripping away more irrelevant data fields such as “production company” as this doesn’t really have an impact on the plot of the movies etc.

So for next week I will:

Begin working on my report

Really start getting into testing properly

Finish up cleaning and find tuning the system

[danielwatts99-blog]

04/03/2021

update

So this last week I have mainly been working on some data science techniques as well as some visualisation, so I have added a small trends part to my system that tries to find any correlation between movies and then what gets recommended, for example I have created a chart that compares the movies with the highest budget and see where they end up amongst the highest revenue generating and most popular movies.

For next week I will

* Improve upon the data analysis segement

*work on improving the accuracy score of my recommendations

[danielwatts99-blog]

25/02/2021

This week I spent much of my time cleaning the code and trimming down the unnecessary components such as rewriting functions and just simply calling them instead, adding more markdown text so it does not look like a long script of code but more of a report style. I do want to add more graphs and charts as well.

For next week I will

*Add more data science techniques

*Ask Tim about any features he think could be added

*Start plan for draft report

[danielwatts99-blog]

18/02/2020

Update

This week I have focused on the user based filtering part of my project, which takes a user that is searched for by their id, then a movie id, this Is cross referenced with users who have liked similar movies and what the predicted rating would be. I have found this extremely difficult as the library Surprise I was using no longer supported the evaluate function which was key for this part of this segment, so after MANY hours of research I then found out that there was a similar function called cross_validate, so I had to alter the code in order to support this, which then completed this segment of the recommender system.

For next week I will

*clean up the code, it is very messy at the minute and makes debugging and troubleshooting unnecessarily difficult.

*Look for additional features to add, I want to add some data science techniques as well for example finding certain correlations and trends

*Begin testing plan/I have decided to go with user stories and test cases as my method of testing, similar to how app developers test.

[danielwatts99-blog]

11/02/2020

Hi,

So after my meeting with Tim last Friday we discussed a new way to write these posts, so now I will be writing these posts in a similar way to agile working sprint reviews, highlighting what I completed the last week, what worked well, the things I learned and what I am going to do in the future.

So from the last meeting I have decided to work on a search function that the user inputs to find a specific movie, the recommendation part comes later but for now I am just focusing on the user inputting a movie and then running it through the database to find a result or a correspoddence, this has proved really tricky as some of the code I have produced is so specific to the letter and number and caps, so this is something I am going to perfect by next week.

Overall I feel like now the project is really starting to take off and the implementation is running quite smooth with no real tough setbacks as of yet, of course there are hurdles I have overcome but this is to be expected.

[danielwatts99-blog]

Christmas break

Hi,

I have had a very busy christmas period, mainly progressing my prototype for my report. The report is a thousand words summarising what I have done so far and my plans for the developing and testing including a timeline with detailed milestones. There is still a lot of work to do but I feel that i have chosen a dataset that allows me to keep expanding and expanding.