We wanted a web-based project and had just recently learned about Twitter and Node.js in class.
What it does
It searches through recent tweets for medical keywords, and either replies with specific advice if available, or links to outside information. The bot also retweets news from major medical accounts.
How we built it
We broke up the bot into different base functions: retweeting, tweeting, searching for tweets, and searching for links/information. Once we did that, we focused on the basics for each of our functions and gradually expanded their utility to interact properly with each other.
Challenges we ran into
We intended to use TensorFlow to create a reverse image search, however support was lacking so we had to abandon that.
Accomplishments that we're proud of
What we learned
We learned that primarily using string parsing to determine replies can lead to some humorous outcomes.
What's next for Medical Advice Twitter Bot
Implementing direct message capabilities, reverse image searching, expanding reply language, and refining the string parsing to filter out some extraneous replies.