I'm taking an NLP class this semester and it's a lot of fun, I've done some work with Twitter bots in the past and I thought this would be a cool thing to make.
What it does
It has two interfaces, a Twitter bot and a web app; for the bot you just need to tweet @ it and mention some other people in the tweet, and it will reply with one tweet it generates from the people you mention, and for the web app you just type in some Twitter handles and specify some parameters and you can generate as many tweets as you want.
How I built it
I built the NLP and Twitter bits in Python because Python has a nice Twitter API library that I'm familiar with, and then I made a Node.js server to host the web app.
Challenges I ran into
At one point there was something wrong and I was very tired and couldn't figure it out and turns out it was because one line I had intended to produce floating point values was being rounded to integers.
Accomplishments that I'm proud of
I'm pretty proud of it overall, it's fun to use
What I learned
I learned how to use Pug to build HTML templates that can be filled with data (like the generated tweets)
What's next for Tweet Fusion
I might rework the N-gram models to use some cool association measures we learned about in class last week