Inspiration
I was watching a horse race before the Datathon started, so I decided to go with this.
What it does
It loads the profiles and race histories of certain horses (they all fall under a certain category-- I wonder if you can figure it out?) and their stats. You can also predict the chances of a horse winning for a given race, or ask the app a question and get an SQL query as an answer.
How we built it
We web scraped Wikipedia for a bunch of data on horses (many other APIs were not free, so we decided to just use Wikipedia) and built a machine learning model on the race history.
Challenges we ran into
Finding and fixing unexpected bugs was the biggest challenge. For example, we had failed to account for bounds checking for a certain feature which threw many things off.
Accomplishments that we're proud of
Building an app that can process natural language into an SQL query
What we learned
How amazing and easy AI is.
What's next for Horse Racing Data
If we were to continue this project, it would be getting access to more data using better API's and adding features that would help users better understand what they are looking at.
