Twitter Harassment Labeling Tool
This is the repository for the tool hosted at http://twitterharassmenttool.tk
To help label tweets as harassing, or not harassing, please visit http://twitterharassmenttool.tk
To get started running the project, clone the repo,
npm install then run
The repository includes around 8000 starter tweets in
tweet_data/0.tweets.dat. These tweets were gathered from Twitters random tweet sample feed over a period of around 8 hours. The format for tweet data files is newline delimited JSON objects for each tweet. Each tweet has a string tweet id, the number of ratings that it has been given, and the average rating. The starter tweets are unlabeled.
Please open a PR if you would like to improve this tool. I'm pretty flexible on contribution guidelines, but as a general rule, if in doubt, keep it short, and make it count.
Everything is GPL3+, so feel free to host your own version elsewhere if you want to gather data for a different set of tweets.
This isn't asscociated with Twitter in any way, and complies with all aspects of the Twitter developer agreement. Thank you for your time.