/pitchfork-data

Analyses on over 18,000 pitchfork reviews.

Primary LanguageJupyter Notebook

Pitchfork Data

I scraped over 18,000 Pitchfork reviews, going back to January 1999. I'll be analyzing the data to satisfy a few of my own curiosities.

Notebooks:

  • [Link]. Some evidence of statistical heaping in the review scores.
  • [Link]. An exploration of review scores.
  • [Link]. Do writers get tougher with experience? (answer: no.)
  • [Link]. Is "Best New Music" sampled IID? (answer: yes.)
  • [Link]. Is the first album the best? (answer: no, but the last album is the worst.)
  • [Link]. Are reviews autocorrelated? (answer: yes.)

Some other things i want to know:

  • Are best new music reviews longer? shorter? Same Q for very low ratings.
  • What is avg. amount of time between original and reissue date?
  • Can candidates for re-release be predicted based on first-round review content?
  • Can you predict when an album should be re-released based on current data?