Week5

Week #5 #

Feedbacks: #

Feedback collection plan

We decided to find people who are interested in mashups genre, give them our product and collect all the problems they encountered as well as all the suggestions for improvement.

Conducted user surveys or feedback sessions

Ten students of different universities, including Innopolis University, were asked to try our product and describe the problems with user experience they encountered in the process of using the product.

Analyzing feedback, identifying and prioritizing issues

Feedback showed multiple issues among which we chose those that affect the most basic user experience to focus on fixing them first. Among them are:

  • Crossover feature is unintuitive
  • Player does not indicate the exact track from playlist being played
  • Playlist page has a very long loading time
  • There are no any hotkeys
  • There is no opportunity to see all the mashups of a certain author
  • There are several visual bugs which make understanding the pages difficult

Roadmap: #

  • End of July - add mashup uploading functionality
  • First half of August - add functions to import mashups from VK and create collections of mashups for recommendation
  • Second half of August - add mashup ‘radio’ (endless stream of mashup tracks) and Discord integration
  • Fall - create mobile application

Weekly Progress Report: #

  1. Frontend:
  • User can now listen to playlists
  • Player page added
  • User authorization implemented
  • Backend mostly connected to frontend
  • Several fixes issued
  • Almost finished the project
  1. Backend:
  • Optimization improved
  • More tests created

Challenges & Solutions #

We were faced with too small amount of existing features in our current dataset, so ML model may be not very good. We still think how we can extend them to improve quality of our model.

Conclusions & Next Steps #

We plan to finish the project, which includes:

  • Fix more problems with user experience
  • Implement more suggested features
  • Connect all backend with frontend
  • Finish few non-implemented backend routes
  • Improve ML model