Data analysis of the YouTube "PointCrow Clips" channel illustrating engagement over time relative to community efforts to get Lauren (editor, channel manager) a raise.
Only data publicly-available over the YouTube APIs is used.
- YouTube API Key
- This is a good guide: https://www.geeksforgeeks.org/youtube-data-api-set-1/
- In this repo, the key is assumed to be the environment variable
YT_API_KEY
- Python3 requirements
pip install -r requirements.txt
- Recommended using
virtualenv
data/
- Output data from the scripts being run in this repo
script/
- Python scripts which pull data from the YT API
TL;DR
- The channel ID is
UC6MnY4d56I8j2H327vZzmFg
python3 ./script/get_channel_id
- Output stored in
data/videos/list.json
- Raw output stored in
data/comments/raw/comment_threads.json
- Output stored in
data/comments/agg/comments.csv
SQL scripts stored in sql/
. DB is assumed to be PostgreSQL.
- total comments
- Total parent comments
- Total threads with replies
- What are the most-liked comments?
- What is the most-commented video?
- Which comment thread has the most engagement?
- Which video is most disliked?
- Any correlation between duration and performance?
- When did the first raise-related comment appear?
- Before or after the initial clip?
- Did anyone retroactively comment about the raise on older clips?
- Who has commented the most about the raise?
- Did the raise produce a noticeable increase in comment-based engagement?
- Same as raise, but for getting laid