Home

Welcome to the LAVA wiki!

What is LAVA?

LoL Automated Video Analyser (LAVA) is a masters thesis-turned-side project in League of Legends analytics. The idea behind the tool is to automate the process of analysing player location data in VoDs. That is to say, it gathers the positions of all 10 players for every second of a video replay and tries to offer some basic insights on the data, while also saving the data for more nuanced studies.

The goal was to automate as much of this process as possible, meaning this is easier to maintain and use than other projects. There is minimal machine learning/AI involved, so there are no complicated models to train, no training data to gather and updating is as easy as updating your assets folder to include new champions. Graphing, identifying the champions, saving the data, reading the video from Youtube URLs/playlists; all of this and more is done automatically.

It's an open source project, meaning every line of code is available to see (and changes are welcome). If you have any questions/requests, feel free to reach out or raise a PR/discussion on this repository

For more detailed information, find the relevant section here:

• Install
• Usage
• Output
• Issues
• GPU Support/Lightweight Mode