/somafm_charts

Simple scraper to collect and analyze the "Top 30" charts from SomaFM.

Primary LanguagePythonMIT LicenseMIT

SomaFM "Top 30" Charts

This simple scraper parses and compiles the "Top 30" charts at SomaFM. SomaFM is "listener-supported, commercial-free, underground/alternative radio" based out of San Francisco. Their 30+ uniquely flavored channels are hand curated and cover everything from downtempo beats to indie pop rock to metal to classic soul.

Background

Every week SomaFM publishes charts for each channel with the 30 most popular artists, albums, and tracks for the previous week. Each channel has a new page for the week's chart. Here is an example chart page. This tool scrapes the following charts:

  • Top 30 Artists by Spins
  • Top 30 Albums by Spins
  • Top 30 Tracks by Spins
  • Top 30 Artists by Listeners
  • Top 30 Albums by Listeners
  • Top 30 Tracks by Listeners

Definitions from the SomaFM website:

Charts are listed by spin (the number of times we played the song that week) and by listeners, which is the sum of the number of people listening to a channel each time we played that song or artist.

get_somafm_charts.py

This tool goes through each channel's entire available chart history and currently creates a CSV for each channel. The end goal is to push this info to a database.

The chart data is reshaped so each record contains the following fields: week, rank, artist, media, media_type, score_type, score, station, url

To get the full "Top 30" dataset the code has to run through roughly 10 years of weekly pages for each of the 35 channels. It takes a couple hours on my machine to finish.

Dependencies

Developed with Python 3.6 on Windows 10.

Todo List

  • Basics
    • Pull data from all stations
    • Pull data from all dates (programmatic 7 day delta)
    • Parse chart into columns (csv style)
    • Output CSV
    • Args for date range, stations (other filtering after data pull)
    • Start date to 'last Saturday' (relative to today)
    • Modularize for CLI usage
  • Reporting
    • Generate metrics (top artist on groovesalad in 2015, top song on xmas ever)
    • Generate visualization
    • Generate year-end ranking charts, etc.
  • Databasing
    • Create database architecture
    • Save data in database

License

MIT License © Aaron Penne

Disclaimer: This project is not affiliated with SomaFM in any way.