In this post we will discuss λ Lovelace’s progress on the following:

  • iOS app upload user feedback
  • iOS evaluation app submit test result data
  • Challenges & Next Tasks

iOS app upload user feedback

Our recommender system are able present user most possible interested tweets at the top of the list, which is inferred from users’ history tweets, retweets and likes. Mostly it works fine, but what if user want to tweak recommender system by themselves, for example he is not as interested in tweets from Google official account as what recommender system think. We’ve add swipe buttons in home line of iOS client where user could tell recommender system his preference.

Till now, we finally implemented uploading feedback data to Flask server where they are stored in our database eventually. Following is how user preference feedback looks like in our database:

iOS evaluation app submit test result data

We polished evaluation app to be ready for this week practical evaluation test. There are two major work we’ve done for evaluation app. First, tweets displayed in evaluation app are fetched from original tweet homeline list or our recommender system list randomly. Second, we finally implement save test results in database for future analysing. Following is how evaluation test data looks like in our databse:

Except these two major work, we also implement functions such as restart new test, user logout. In conclution, we’ve done all prearation work and be ready to process evaluation test. We also designed a flier for recruiting participants.

Cooperate together

Most work we’ve done this two weeks is establishing interfaces between seperate modules, for exampel we built channels between ios clients to flask, flaks to recommender system and flask to database. We are proud of what we’ve done and enjoy this experience cause these tasks couldn’t implement by anyone of us instead they need all of us cooperate together.

Challenges & Next Tasks

  • Recommender System: Recommender System will involve user feedback data in weight calculation for each tweet and will also implement recommend unfoller tweets this week.
  • iOS: iOS client will implement video and image preview function and add user logout function.
  • All: All of us will clean up our code, add essential comments, do necessary refactor. Meanwhile, we will prepare the final week presentation.

Until next time, on behalf of λ Lovelace.

- Junyang Ma