Bradley-Terry Rankings for Global League of Legends

Inspiration

Hi Rioters and Amazon Devs, we are recent graduates that love League of Legends. We were super excited to participate in this hackathon to leverage our computer science minors and work experience with code. Really loved the challenge of ranking worldwide teams, and learned a lot throughout this process!

What it does

We built a ranking system using the Bradley-Terry Model, which allows us to take a table of wins and losses for each team, and outputs a ranking based on the probability of each team winning against each other. The higher the probability, the higher the ranking!

Challenges we ran into

Most notably, we ran into the "Internal Server Error" when trying to set up the GET request REST API to set up a reachable link (for submission basically). Other than that, challenges came with learning AWS for the first time, trying to use s3, Athena, Lambda and API Gateway. We even tried our hand at sagemaker, but it was a little much for us :)

Accomplishments that we're proud of

Proud of the growth and learning throughout this hackathon. I believe we were super close to getting a final product in but fell short at the end. Testing our code with various parameters works on Lambda, but couldn't get the frontend REST API part to display a valid response.