1. Abstract

Arrigo is Driblab's latest innovative solution, seamlessly integrating event and tracking data to offer a robust set of metrics. Purpose-built to enhance data-driven decision-making, Arrigo delivers deep insights into player actions, team dynamics, and strategic execution. Covering critical areas of gameplay Arrigo empowers teams to analyze real-time strategies, identify performance patterns, and make data-informed adjustments for optimized outcomes on the pitch.

2. Match Preprocessing

2.1.1 Data Retrieval

Tracking Data

• The Driblab API is used to obtain an AWS link, enabling the download of the tracking JSONL file.

Event Data

• Event data is retrieved through the Wyscout API, capturing detailed in-game events.

Offsets

• Two types of offsets are applied to synchronize data effectively:
  1. Footage Offsets: Derived from video footage, capturing ball and player detections.
  2. Event Offsets: Obtained from Wyscout to align accurately with the events data.
• We calculate the Offset Diff, representing the difference between these offsets across game halves (including both extra-time halves). This difference is essential for synchronizing tracking detections with event data.

Data Cleaning and Preparation

• The initial line in the tracking data, containing plyr_lookup.json, is extracted and removed.
• Key pickle files are created for consistent player and team identification:
  • wy_id2plyr_id.pkl: Maps players' Wyscout IDs to Driblab's player IDs for unified analysis across events and tracking data.
  • wy_id2team_id.pkl: Provides similar mapping for team IDs.
• Additional information is saved in match_info.json, including:
  • side_attack
  • home_color and away_color
  • Aliases
  • Other relevant match details.