Time Log

[lwu1822]

Total time: 36 hours

Time: 5 hours

• Researched precision-recall curves
• Adapted code to music recommender to see program's accuracy

Time: 5 hours

• Created frontend to recommend songs (takes user input)
• Created frontend to recommend songs based on a few selected user songs
• Updated navbar to show user login/logout

Time: 2 hours

• Encountered numerous bugs because of the use of sklearn, which caused package version issues. Had to revise requirements.txt multiple times.

Time: 1 hour

• Converted the song recommender to an API endpoint

Time: 2 hours

• Created a song recommender, learned about collaborative filtering from a video
• Played around with .csv files, had to remove a lot of data because the files were too big to commit to Github

Time: 1 hour

• Created an API endpoint that provides a Spotify link to the music
  • Needed to use bash scripting again to filter data

Time: 2 hours

• Created login
• Created sign up
• Created an option to input country
• Created a profile page that shows recommended music based on user's location

Time: 30 minutes

• Created an endpoint for the music recommender
• Created fetch request on frontend for music from the recommender

Time: 1.5 hours:

• Relearned bash scripting to filter data, original dataset came with music id, artist names, track names, streams, etc., needed to filter dataset to only include an id and streams. Note: There probably are other ways to filter the dataset but bash scripting came to mind first.

Time: 4 hours:

• Finding datasets with country and music preference data
• Found a dataset using Spotify charts with all the countries
• Adapted countries data to original artist recommender data

Time: 13 hours:

• JWT - 6 hours
• Debugging JWT - 2 hours
• Finding music datasets - 3 hours
  • Looked through Kaggle datasets, Million songs dataset, and last.fm datasets
• Alternating least squares (used a Youtube tutorial) - 2 hours