Inspiration
We're nerds and we love numbers. We're also voracious users of Facebook Messenger with around 200k messages in the service each, we've accumulated a lot of data (one might even call it BIG data). For us, understanding our Messenger habits would give us huge insights into our communication patterns.
What it does
Our app places a little chat bot inside whatever group chats the user chooses. It then provides a dashboard for that chat with analytics on their activity. This dashboard includes data on any potential cyberbullying going on in the chat.
Tech Stack
The bot and the analytics are both written in node.js, hosted on an AWS EC2 instance. Each group chat thread is stored as an item in a table in Amazon DyanamoDB.
Challenges we ran into
We don't have any super experience Javascript developers and this was all of our first time's with node.js so we struggled a lot to get to know node.js. One of our team members was especially frustrated by Javascript's policy of never failing and never crashing.
Accomplishments that we're proud of
It's neat: it does a good job of telling us about our own chats. Additionally, the Natural Language Processing that it does to analyze bullying does a good job of recognizing aggressive behavior. It successfully differentiates between passive aggressive comments like "..." and fully aggressive comments like "f*** you". Additionally, we're pleased with the pluggability of the database - it'd be super trivial to swap out another chart.
What we learned
The aforementioned teammate that complained about Javascript's lack of exceptions and pauses has learned to adore JavaScript. We also learned a lot about the traditional frontend-backend-database structure and how to make sure they work well together.
Log in or sign up for Devpost to join the conversation.