VALORANT-SSAC22
Code and Data used in Winning Duels in VALORANT, finalist for Sloan Sports Analytics Conference 2022. Vote for this paper if it's your favourite in the open source category here!
The PAPER.pdf and POSTER.pdf are final, but the repo will continue to be updated over with more data and better documentation.
Current files are notebooks which generated the plots used in the paper, added for reproducibility purposes.
Created using Python 3.8.8, Anaconda installation with a few other libraries like xgboost.
FAQs
-
How did you get involved with NRG? I started by making heatmaps on the Valorant Competitive Subreddit and the coach of NRG JoshRT followed me on Twitter. I sent him some ideas, and luckily for me he was willing to try it! Data only takes you so far, seeing it in action has been very rewarding.
-
I saw your heatmaps on reddit, how is this different? The heatmaps were meant to be a visualiation tool of x and y positions only. It does not account for the context of a fight such as guns and armor, that's where the model shines.
-
Where is this data coming from? The data is provided by the VALORANT Data API, and is sorted and parsed by Runitback.gg. None of this would have been possible without RunItBack's support. Join their discord, it's a great place to start with VALORANT analytics.