Inspiration
Both of our team members have been playing and watching League of Legends esports for multiple years and were immediately interested in working on this project. We both also have data and coding backgrounds and were excited to combine our passions.
What it does
Our ranking is displayed on a streamlit web application, allowing the user a customized and in-depth look at our ranking system. In the web application users can choose from the tabs team ranking, global ranking, and tournament rankings. In the team ranking tab users can specify a date or date range and choose teams they want to get the ranking of at that time. In the tournament tab, users can select a tournament and receive the post-tournament rankings. Finally, in the global rankings, the user can specify the number of teams and a date to receive the top teams in the world at the specified date up to the number of teams the user asked for. The user may also specify a date range to see top team ratings over time.
How we built it
Challenges we ran into
While working on this project we did run into various roadbumps. First of all, we initially had trouble properly joining datasets together. We generated a table containing the match history and many relevant statistics from the many games files, but were unable to meaningfully connect it to the tournament table as we thought that the tournament a game was taking place in should impact the ranking. For example, worlds finals should be more important than a random regular season game. However, since there was not a direct key we could use, we had to think of other ways such as connecting based on match day falling within a tournament's date range and looking at which teams played. We also ran into issues when it came to ranking minor region teams and LPL teams. Minor region teams that dominated their respective regions, won many games in playins then were eliminated in one qualifying bo5 versus EU for example would overperform in our rankings initially. We also had a persistent issue of certain LPL teams such as RNG and IG being overrated despite performing poorly in 2023. We eventually realized this was due to missing historic LPL data, which was later uploaded to the S3 containing all the other data. We then had to figure out how to incorporate this into our ranking despite missing most in-game stats and some other variables.
Accomplishments that we're proud of
We are proud of creating a working ranking out of this difficult-to-work-with data and building a working web-app using streamlit.
What we learned
We learned a lot about using AWS Athena and how to query complex data into much more usable formats. We also learned about various elo rating systems and how to modify them to fit our needs.
What's next for ForgeLol
We will look to continue working on our project and touching it up as it is something we are both immensely proud of.
Log in or sign up for Devpost to join the conversation.