The data I used for this project was collected and is continuouly being updated my Tim Sevenhuysen of Oracles Elixir. The site contains data about competitive League of Legends (LoL) across different regions and also international events. The regional leagues include the North American LoL Championship Series (NALCS), European LoL Championship Series (EULCS), LoL Champions Korea (LCK), LoL Professional League (LPL – China), LoL Master Series (LMS – Taiwan) and Circuito Brasileiro de LoL (CBLoL). The site includes data from the 2016 Spring Split up to the current ongoing games so the dataset is currently being updated to the latest matches finished. As of the time of writing, Mid-Season Invitational 2017 just concluded its Group Stages and the data from MSI can be found on the site.
For this project, I used the 2017 Spring Split Regular Season data, and I chose the data from the "main" regions: NALCS, EULCS, LCK, LPL and LMS. Each row in the dataset contains individual player's game stats and some rows contain overall team game stats. I focused on looking at the champion picks, bans and performance through the different patches. The regular season was played on patches 7.01-7.06.
Obviously, I play League of Legends and I follow competitive games a lot. I'm a relatively new player since I started playing just before Season 5 started, and I was never good the game (Here's proof! Currently, hard-stuck in Bronze 2 with over 500 ranked games played this season). I've always wanted to work on a side project related to league but I haven't had the chance for it so I used this opportunity to finally make something about league! There's a lot of things on the dataset that I'm using but this time I focused on how the picks and bans change throughout the patches (and should Riot nerf Camille to oblivion?).
I spent a lot of time moving my data around to fit my needs and I had to filter out some regions (sorry, Brazil). I have used Trifacta, Java just because I had Eclipse open when I started this, and R to process my data. The columns I used are the patch number, bans 1-5, picks, role, win/loss, region and of course, the champion name. These are mostly categorical data so I had to produce counts of champion bans, wins, and picks. Further explanation of these would be indicated in the visualizations.
That's all for the description of my project! You can click here to check out my visualizations.