PredictIt vs. 538 is a Python / Flask data-based web app that compares election predictions from two popular websites for the 2020 US presidential elections.
This was my final project for my CS50 class at Harvard. As someone who checked both of these websites quite often in the months leading up to the election, I was interested in the results based on "gut feeling" vs. statistical analysis.
I am especially proud of this project because I dreamed it up myself and figured out how to implement it all on my own. To accomplish this, I had to figure out what an API is, and how to use JSON, how to host a Python-based, dynamic site online, along with other essential web development skills.
The result was a site that calculated up-to-the-moment stats and that fulfilled a practical need.
Fortunately, the 2020 election is now behind us, but unfortunately that means that the website no longer updates itself. However you can still check it out here.
538 stopped updating its predictions the night before the election, and PredictIt continued updating for months afterward, so the numbers on the site don't make much sense anymore. However, you can navigate around and see how the site works, the different prediction markets I was keeping track of, and get the user experience.
For a glimpse of the site back when it was totally live, check out this short video walkthrough: