spotify-data
├── Read_Me_First.pdf         # Introductory document
├── Follow.json               # Data about user's followers
├── Identifiers.json          # Identification information (🙅 Do Not Share!)
├── Identity.json             # User identity details (🙅 Do Not Share!)
├── Inferences.json           # Inferred data from user activity
├── Marquee.json              # Marquee-related information
├── Payments.json             # Payment details (🙅 Do Not Share!)
├── Playlist1.json            # Details about a specific playlist
├── SearchQueries.json        # User's search queries
├── StreamingHistory0.json    # User's Streaming history part 1
├── StreamingHistory1.json    # User's Streaming history part 2
├── Userdata.json             # General user data (🙅 Do Not Share!)
└── YourLibrary.json          # User's Spotify library details like likes, albums, artists, etc.

Philosophy

I want to analyze my Spotify’s Streaming History in a way from which I can know my listening pattern over the time. The way Spotify give the Spotify Wrapped at the end of every year.

The Dashboard

There has been a dashboard (using Streamlit) where other users can upload their StreamingHistory.json files to see analysis on their history.

Some Awesome Insights

To gather insights from json files I’ve used polars and their builtin .plot accessor which uses hvplot library under-the-hood.

See the Jupyter Notebooks to see all the insights.

Using StreamingHistory.json

Contains User’s Streaming History with trackName (Streaming Track Name), artistName (Streaming Artist Name), msPlayed (Milliseconds Played) and endTime (When the track ends (as datetime)).

• Top T/As in whole dataset.
• Top T/As in each month.
• Monthly most listened Tracks and Artists.
• First day when T/A was played.
• No. of distinct T/As listened in each month/year.
• A T/A streaming in bar plot (which shows how you stream that during time-to-time).
• Which daytime user listen most and whom.
• Tracks which have listened most times in a day.
• Tracks streaming streak (by day/week).
• T/As which only played once
• Dates when user does not played any track

Using Playlist.json

Contains User’s Created Playlist Data with all the tracks added in the playlist.

• No. of T/As and Albums in each Playlist
• Playlist MinutesPlayed
• Most streamed P/As
• Check any Track present in multiple Playlists
• Streaming timeline of P/As (with plot.line())
• Playlist-wise top T/As

Using YourLibrary.json

Using Marquee.json