DetectBot aims to generalize bot detection across social media platforms. It uses a dataset of ~2800 twitter accounts to train a decision tree classifier.

Depending on the type of social media provided, DetectBot retrieves the necessary features from that user's profile and then makes a prediction using the classifier.

This is mostly a proof of concept for me to learn more about ML, and is not meant to be a serious machine learning project.

Usage

go run main.go -c=config.yaml -url=https://twitter.com/username

Config

You will need to apply for a twitter developer account and create a config file. You can pass DetectBot the config using the -c flag.

twitter:
  consumerKey: ""
  consumerSecret: ""
  accessToken: ""
  accessSecret: ""

Supported Social Media

Twitter User Profiles

TODO

Twitter Threads
Facebook/Reddit/Instagram User Profiles and Threads

License

GNU GENERAL PUBLIC LICENSE