This repository contains Twitter data, in csv and json formats, made up of 36k tweets and 17 features. The dataset was created to build a classification model to help classify the emotion of a Tweet as anger, joy, fear, or sadness, and then output this classification as an Emoji.
Future work for this project includes building (or acquiring) a corpus of text with content classified as "violent", and work on detecting violent or potentially violent Tweets for flagging.
A presentation detailing the project is available at: https://docs.google.com/presentation/d/1BF_v7UviBcaGRjsOArbECirm8xn_c-F12XtXv-mAvSc.
This project was submitted as part of my masters pattern recognition coursework.