Skip to content
Back to projects

cancel culture impact

Measuring whether celebrity cancellations actually hurt careers

PublishedJune 2024
Open source
Python
LibreTranslate
Statistics
Data Science
Problem
When a celebrity gets "cancelled", does it actually hurt their career, or does the noise just move on?
What I built
A data pipeline that crawls news, YouTube comments, Spotify metrics and Billboard charts around four cancellation events, then tests sentiment change statistically.
Role
Solo project
Stack
Python, GNews API, YouTube API, Spotify API, VADER, Jupyter
Outcome
A measurable impact in almost every channel, for all four cases studied.

Does cancellation show up in the data?

Short answer: mostly yes. I picked four heavily covered cancellation events (Kanye West, Seungri, R.Kelly, Marilyn Manson) and checked each data source for a measurable change around the cancellation date. The table marks where a significant shift showed up.

Impact metrics based on data analysis
Kanye WestSeungriR.KellyMarilyn Manson
YouTube
Articles
Spotify

Article sentiment shifted in every case. Audience metrics (YouTube, Spotify) moved in most cases too, but not all: R.Kelly's YouTube data and Marilyn Manson's Spotify numbers showed no measurable change.

How I measured it

Everything is Python: preprocessing, then sentiment scoring with the VADER lexicon after translation, stopword removal and lemmatization. To claim a real effect rather than noise, I ran paired t-tests comparing sentiment before and after each cancellation date.

The crawlers

Four custom crawlers feed the analysis: GNews for articles, the YouTube API for comments and video stats, Spotify for daily follower and listener metrics, and Billboard charts from 2015 onward for chart performance.