Context

I wanted to study player stats for Australian Rules Football (AFL), using Machine Learning to identify who the key players are, predict performance and results. My aim is to create the Ultimate Tipping predictor or maybe a Fantasy League tool. I couldn't find a dataset anywhere so I created my own GCD project and am sharing the database for anybody to explore.

Content

Every key stat from Kicks to Clangers to Bounces. Every player, game by game.

Updates in the latest version:

• Seasons 2012 to 2024
• Progress scores at quarter time, half time and 3 quarter time breaks
• Max & Min temperature on game day

Joining Data

Primary Keys are unique values in 'PlayerId' and 'GameId'. Do not join by 'DisplayName' as these are not guaranteed to be unique. 'Year' and 'Round' columns are duplicated across stats.csv and games.csv for convenience when reading.

AFL Domain Notes

• Docklands is the only stadium with a roof that might be closed on game day
• Season 2024 onwards has 'Opening Round' (OR) as the first round of the regular season, with only some teams scheduled to play. This is like a 'Round 0' as the AFL calls the following week Round 1.

Future Versions

If you have any further suggestions please Comment below.

Acknowledgements

Data is taken with thanks from afltables.com and www.footywire.com

Inspiration

I want to see key insights into the players' performance that nobody has realised before. With tipping contests in the AFL as popular as any other sport, surely this is just waiting for Data Science to take over and pick winners like never before!!