RedFlaggerBot

Using Heroku to operate our suicide and depression detection bot on cloud, we seek to help parents be more aware of their child's mental health, by mentoring their social media posts.

Comment

Inspiration

Having gone through the painful experience of losing our friends to suicides and depression, our team wanted to prevent friends and families of potential victims from such unfortunate events. The number of suicides in Singapore rose 10 per cent last year, with suicides among boys aged 10 to 19 at a record high. Inspired to address this societal concern, we ventured to find a way to sieve out potential victims, so they can receive help promptly. Social media platforms are commonly used for the expression of suicidal thoughts and feelings, particularly by young people. Hence, we decided to tap on this open platform to help save the lives of youths in need.

What it does

Parents can monitor their child's emotional state simply just by entering their child's Twitter username. Our bot will sieve out their posts and run through algorithms and check it against our database of keywords to rate their emotional health. This will give parents a good sense of how well their child is doing. For critical cases, we will flag them out and advise parents to seek help immediately.

How we built it

  • We each segregated our parts into Twitter Integration & Web Scraping/ Hosting on Heroku & Retrieving/Interacting with Telegram bot.
  • Using lambda & python to retrieve basic user inputs and commands to stimulate basic interaction with the user and also to retrieve the information needed for web scraping.

Challenges we ran into

  • Hosting our bot on Heroku as it was our first time doing so. We just started learning about its functions hence we were unaware of how to troubleshoot our errors. In the end, we managed to debug and found the mistakes, and we were able to successfully host our bot online.
  • We had difficulties determining the parameters and data we were going to capture, and linking them to carry out the checks.

Accomplishments that we're proud of

  • We managed to create and test out the bot real-life within the stipulated time

What we learned

  • How to host a bot online on Heroku, which we faced a lot of trouble with at first
  • Logical flow analysis was vital in determining what metrics we were going to use for Firebase

What's next for RedFlaggerBot

  • Increase the scope of social media sites to Instagram, Facebook
  • Be more specific by flagging out the cause of depression (e.g. personal relationships, academic stress etc)
Share this project:

Updates