Making a Chrissy Teigen Bot - Summer Projects
In addition to organizing the March for Science Student Summit (shameless plug - more information here), I am trying my hand at a new machine learning project! I’ve never worked with natural language processing before, and wanted to learn about data mining, so I decided to make a program that creates tweets based on training data from a particular user.
Initially, I wanted to use myself, but I needed a large enough dataset to even have a shot at generative realistic tweets. I also needed a user that was not likely to become… problematic (a la Microsoft’s super racist twitter bot). After a thorough search (read: I opened Twitter and scrolled twice), I found the perfect person - Chrissy Teigen!
So far, I’ve downloaded the tweets using the Tweepy API and done some light pre-processing to remove unncesessary metadata. I’m planning to pick a model over the next couple days, and will update this post as things happen!
You can follow my progress on GitHub here: https://github.com/harrodjordan/ItsChrissy