Last.fm Listening Autobiography #

A toolkit for exploring your Last.fm listening history, cross-referencing it with your Discogs vinyl collection, and building a biographical context layer that turns scrobble data into actual stories.

Built in conversation between Joel Goodman and Claude. The code is generic — the stories are yours.

What's in here #

lastfm_client.py         # Last.fm API wrapper — all user-relevant endpoints
context_generator.py     # Formats data into injectable context blocks
lastfm_stats.py          # CLI entry point — 13 commands
listening_bio.py         # Biographical layer (YOUR stories go here)
SKILL.md                 # Claude skill definition for integration
bio-interview-questions-template.md  # Question template for the biographical session

Quick start #

1. Get a Last.fm API key: https://www.last.fm/api/account/create
2. Set environment variables:

   export LASTFM_API_KEY="your_key_here"
   export LASTFM_USER="your_username"
   

3. Create a virtual environment and install requests:

   python3 -m venv .venv
   .venv/bin/pip install requests
   

4. Run it:

   .venv/bin/python lastfm_stats.py profile
   .venv/bin/python lastfm_stats.py snapshot --period 7day
   .venv/bin/python lastfm_stats.py years
   .venv/bin/python lastfm_stats.py year 2015
   .venv/bin/python lastfm_stats.py genres --period 1month
   .venv/bin/python lastfm_stats.py evolution
   

The process #

Phase 1: Let the data talk #

Run the commands. Look at your year-by-year breakdown (years), your all-time top 50 (snapshot --period overall --limit 50), and the genre evolution (genres across different periods). The data will surface patterns you didn't know about — volume spikes and cliffs, artists that appeared and vanished, genre drifts.

Phase 2: Ask the questions #

Use the interview template (bio-interview-questions-template.md) or better yet, have Claude generate custom questions from YOUR data. The pattern is:

1. Share your years output and snapshot --period overall output with Claude
2. Ask Claude to identify unexplained patterns (volume drops, surprise artists, genre pivots)
3. Claude generates questions specific to your data
4. You answer conversationally — every correction and story gets logged

Phase 3: Build the biographical layer #

Edit listening_bio.py to add your own eras, artist bios, album stories, key people, and live shows. The data structures are simple:

• Eras — life chapters with dates, locations, and sonic palettes
• Artist bios — relationship type (formational/sustained/returning/discovery), era connection, sonic role, personal story
• Album bios — same structure for specific albums
• Pipeline people — who introduced you to what
• Live shows — experiences that shaped the relationship

Phase 4: Cross-reference with Discogs (optional) #

If you have a Discogs collection, the vinyl cross-reference reveals the "blind spot" — artists you care about deeply but whose scrobble counts underrepresent the actual relationship because you listen on vinyl.

# Discogs collection is public by default — no auth needed for read
import requests
resp = requests.get(
    'https://api.discogs.com/users/YOUR_USERNAME/collection/folders/0/releases?per_page=100',
    headers={'User-Agent': 'YourApp/1.0'}
)

Phase 5: Build the timeline (optional) #

The toolkit includes a timeline web page generator that pulls album art from the Last.fm API and displays your listening autobiography as a visual narrative.

CLI commands #

Command	What it does
profile	Account stats
now	Currently playing
snapshot	Top artists/albums/tracks for a period
recent	Last N scrobbles
genres	Weighted genre/mood profile
artist "Name"	Deep dive on one artist
context	Compact block for skill injection
annotated	Snapshot with biographical annotations
bio	Listening biography summary
evolution	Compare across time periods
year 2015	Top artists/albums for a specific year
years	Summary of all years
raw	Raw API JSON output

Integration with Claude skills #

The context command generates a <listening_context> block designed to be injected into other Claude skills. If you use Claude for creative work — writing, visual direction, content strategy — the listening context gives downstream skills awareness of your current sonic palette and biographical relationship to it.

Attribution #

This project uses the Last.fm API for listening data and the Discogs API for collection data. Per the Last.fm API Terms of Service, any public-facing use must credit Last.fm and link back to the appropriate pages.

Album art is served from Last.fm's CDN via their API and is used under their non-commercial license terms.