valorwin-model

Web App Demo

https://valorwin.vercel.app/

Brief Summary

An AI model (now a Full Stack Web App) utilizing random forest classification to assess the likelihood of winning individual rounds and entire matches in Valorant.

Download link for .pkl and .csv files: https://drive.google.com/drive/folders/1rjTZpR41E4Q9DpxqqX8VDf53y9XsvwgK?usp=sharing

Data

• Our data will be scraped from https://www.vlr.gg/
• We have been given permission to scrape their website.
• This dataset has been created with data from over 10,000 bo3 and bo5 matches from the T1, T2, and collegiate scene.
• There are almost 500,000 rounds of data that cover every map on Valorant.

Example Data

Input Variables

• Team Loadout
• Enemy Loadout
• Team Rounds Won
• Enemy Rounds Won
• Map

Notes

• Eco: 0-5k
• Semi Eco: 5-10k
• Semi Buy: 10-20k
• Full Buy: 20k+

Output Variable

• Match Outcome or
• Round Outcome

Classification Report

How To Use

1. Import dependencies and load model

2. Input Data [Must be in this order!]

3. Predict probabilities

4. Output results

Built With

• Requests
• BeautifulSoup
• Python-DotEnv
• PyMongo
• scikit-learn
• pandas

Contributing

Feel free to submit a pull request or an issue!

License

The MIT License (MIT)