YouTube scammers are everywhere, often leaving comments on personal finance videos using bots. These comments often contain contact information of the scammer. If someone contacts them, they will be promised extremely high returns, but once they make an “investment,” the money is gone for good.
The comment sections of channels like Graham Stephan, Andrei Jikh, Tech Lead, and others are filled with these scammers. I wanted to tackle this issue using code to see if we can make the YouTube community a little safer and prevent some people from falling for these scams.
So what did I do? I used the YouTube Data API to gather a dataset of YouTube comments and manually labeled each comment as a scam (1) or safe (0). Then, I trained an NLP model, which achieved over 90% scam comment classification accuracy. Pretty good!
Then, I write a Python program that finds scam comments and leaves a friendly reply underneath, urging viewers to exercise caution with the scammy comment. This makes the YouTube scammer exposed, and viewers are less likely to fall for the scam.
If you enjoyed the video, please like and subscribe!
My code: n
YouTube Data API Access: /
API Key Guide: y
YouTube API OAuth Access: 8
What is Natural Language Processing (NLP): g
Existing solution to spam problem:
Kill Spam: /
If you are curious how to delete spam comments on YouTube, check out the service above. I have never personally used them and am not sponsored by them, but they seem legit and offer a great service to stop spam comments.