This project displays winning probabilities for a League of Legends game in spectator mode in real-time.

If you want more details you can check out this blogpost.

The ETL uses the Riot Games API to gather match data.

A classifier (random forest) is built using Apache Spark on the data previously collected.

A streaming application reads game data sent in real-time by our client from a Kafka topic, interrogates our previously built model and post the probabilities of each team winning to another topic.

The webservice acts as an intermediary between the client and Kafka.

The client captures game data thanks to the LeagueReplayHook, sends this data to the webservice and displays the probabilities thanks to Overwolf.