A Jupyter notebook that finds the top 50 YouTube influencers in a given sector using the free YouTube Data API. No subscription, no scraping, no third-party services.
Built as an open alternative to tools like Modash and HypeAuditor — which do the same thing but charge a few hundred dollars a month for the privilege.
You pick a sector from a list of 20 (fitness, finance, gaming, food, etc.), point it at your YouTube API key, and run the notebook. It searches YouTube across a set of relevant keywords, pulls channel and video statistics, and produces a ranked list of the top 50 creators in that space.
Each channel gets a Strength Score out of 100, made up of five things:
- Reach (25%) — subscriber count, log-scaled so one mega-channel doesn't skew everything
- Engagement (30%) — average likes and comments relative to average views
- Authenticity (20%) — views per subscriber, which is a decent proxy for fake or dead followers. A channel with 800K subscribers and 4K average views is a red flag, and this catches it.
- Consistency (15%) — how many videos they've posted in the last 90 days
- Momentum (10%) — whether their most recent videos are outperforming their own average
Just a YouTube Data API v3 key. It's free, takes about five minutes to set up, and gives you 10,000 units a day — enough for around 20 runs.
- Head to console.cloud.google.com and create a project
- Go to APIs & Services > Library, find YouTube Data API v3, and enable it
- Under Credentials, create an API key and copy it
- Paste it into Cell 2 of the notebook and you're good to go
git clone https://github.com/YOUR_USERNAME/influencerank.git
cd influencerank
jupyter notebook youtube_influencer_finder.ipynbOr just open it in Google Colab if you don't want to run anything locally.
All dependencies install automatically in the first cell. Once you've set your API key and chosen a sector, run all cells and let it go. The whole thing takes a couple of minutes depending on how many channels it finds.
A ranked table in the notebook, plus three charts saved as PNGs — a bar chart of the top 20 by score, a scatter plot mapping engagement against reach, and a radar breakdown for the top 5 channels. The full dataset also exports to a timestamped CSV.
Fitness, Finance, Tech, Gaming, Beauty, Fashion, Food, Travel, Business, Education, Music, Sustainability, Parenting, Mental Health, Sports, Art, Cars, DIY, Pets, Politics. You can add your own by dropping a new entry into the SECTORS dictionary in Cell 2.
This only covers YouTube. Instagram and TikTok have locked down their APIs to the point where doing this properly isn't feasible without paid access. The authenticity score is a statistical signal, not a forensic audit — it's good at flagging suspicious channels but it's not the same as a full fake-follower analysis.
MIT Licence.