Mental Health Analysis Toolkit

The toolkit formerly known as Vibecheck by Neeraj Alavala, Aaron Graubert, and Sean Walsh

Overview

We started this project with the intent to provide a robust mental health analysis pipeline to handle live social media data. While we were unable to complete this project within the semester, we still wanted to organize our work so that others may continue from it in the future. You can find our final report FIXME.

Structure

• discord: This folder contains the start of a discord scraper, however this project direction was abandoned for various reasons, one of which being that SDCNL (the classifier chosen) was trained on reddit
• reddit: This folder contains code for our Reddit scrapers. Further information can be found in the reddit directory
• SDCNL: This folder contains code for SDCNL. Further information can be found in that directory

Getting Started

To resume work on this project, we recommend that you get the reddit scraper up and running first, so it can build up a volume of live comments while you work on other components. Detailed instructions can be found in the README for the reddit directory.

After your scrapers are running, you can use SDCNL to make calls on comments. SDCNL is built in a docker image and the dockerfile is provided. You can also run our version of the image: agraubert/sdcnl.

Data Access

The following data can be made available upon request:

• Disk images for our data scraping and SDCNL VMs
• Snapshots of our Reddit database
• Buzzcard data obtained from Georgia Tech

To access any of this data, please reach out to Neeraj, Sean, Aaron, or Professor Pu.