I really wanted to build an app for the Hack Harassment initiative. I initially wanted to make a chrome extension that can scrape the current web page a user is on and determine whether posts or messages are harassing and then block them. Unfortunately I was only able the first half of the application where I classify whether text is harassing. To do this I used IBM's Alchemy API to do sentiment analysis on text - which only can determine whether the text is negative or positive. I then used the API to extract the keywords and whether the keywords belonged to certain subcategories ie hate crime, racism, ect. If these categories line up then I classify the text as harassment.
Harassment Classifier
A simple chrome application that takes text such as a tweet and tries to identify whether it is harassment.
Updates
Leave feedback in the comments!
Log in or sign up for Devpost to join the conversation.