New Research on Multidimensional Database of In-Game Player Movements in Gaelic football

The research is about creating a multidimensional database of in-game player movements in Gaelic football using GPS data. The study transforms raw GPS data into a series of player actions, which can be analysed to understand player performance, movement, and team structure. The dataset is designed for use by machine learning researchers, sport scientists, and coaches.

GPS data was collected using wearable micro-GPS sensors (STATSports Apex 10 Hz) during 11 competitive Gaelic Football games between 2019 and 2021. The sensors recorded latitude, longitude, and speed ten times per second. Raw GPS data is also processed into player actions by converting speed values into six action labels: standing, walking, jogging, running, high-intensity running, and sprinting. Consecutive seconds of the same action are aggregated. These actions are then grouped into "events" which are sequences of actions beginning and ending with standing or walking.

The final dataset consists of 159,610 actions, each associated with a player, game, and event. It includes 12 variables such as game ID, player ID, action type, start and end times, duration, and distance. The data is available in CSV and SQL formats. Additionally, the dataset is dimensional, allowing for analysis across various dimensions such as game, player, action type, and duration. This creates data cubes at different levels of granularity, including 1-D, 2-D, 3-D, and 4-D cubes.

The dataset can be used for statistical analysis, machine learning tasks, and educational purposes. It can help in understanding player running performance, team structure, and identifying patterns in player behavior. The action dataset was then compared with existing literature on Gaelic Football and showed consistency in terms of distance covered and movement intensity.

An analysis of the dataset revealed that the average speed during the first half was slightly higher than in the second half. Additionally, the mean duration of high-intensity running was significantly longer in the second half, potentially indicating fatigue. Overall, this research provides a valuable resource for analysing player movement and performance in Gaelic football using a data-driven approach.

Read the full paper here.