The purpose of this analysis to use NLP to identify users who suffer from mental illness via the content they post on Twitter. The analysis uses the CLP Psych 2015 data from Twitter users and seeks to classify users into one of the following three categories: No Mental Health Issues (Control Group), Depression, PTSD.
All code is found in the Jupyter notebooks: FinalProjectPart1.ipynb, FinalProjectPart2.ipynb, FinalProjectPart3.ipynb in order chronological order of updates/improvements.
Data for the project was taken from the CLP Psych Shared Task 2015 Data set. Per IRB protocol, the data is not housed here to maintain user privacy.
All code, writing, and analysis in this repository is done by Kripa Agarwal ([email protected]) and Ashley John ([email protected]) for academic purposes at Cornell University in Spring 2018.