Permalink
Find file
Fetching contributors…
Cannot retrieve contributors at this time
34 lines (22 sloc) 1.68 KB

The Simpsons by the Data

Code in support of this post: The Simpsons by the Data

It's a Rails app, but isn't intended to be run as a server. It processes data from Simpsons World, Wikipedia, and IMDb, and populates a PostgreSQL database called simpsons_development. The database contains 4 primary tables: episodes, script_lines, characters, and locations

Instructions

Assumes you have Ruby and PostgreSQL installed

git clone git@github.com:toddwschneider/flim-springfield.git
cd flim-springfield/
createdb simpsons_development
bundle exec rake db:migrate
bundle exec rake import_data
bundle exec rake jobs:work

It takes about 45 minutes to process everything with one worker

Analysis

R code to analyze the data lives in the analysis/ folder

Caveats/areas for improvement

  • I deduped some character names when they're printed in different ways, e.g. "TROY" is the same as "Troy McClure", but I certainly did not dedupe all 6000+ characters that appear in the scripts
  • Similarly I manually assigned genders to the top 320 or so characters, who collectively account for 86% of the show's dialogue
  • I did not dedupe any locations
  • Simpsons World is not available in all countries, so the code might not run depending on where you're located

tab