luxury-lakehouse/obso-pausa-inputs

OBSO/PAUSA Input Data — IDSSE Events + ELASTIC Sync

Input event data and ELASTIC event-tracking synchronization results for OBSO/PAUSA computation — two configs covering DFL match events and their frame-level alignment to 25fps IDSSE tracking data. These inputs feed the Off-Ball Scoring Opportunity (OBSO) and PAUSA pass-timing pipelines.

Part of the (Right! Luxury!) Lakehouse soccer analytics platform.

Quick Start

from datasets import load_dataset

# Load events
events = load_dataset("luxury-lakehouse/obso-pausa-inputs", "events")
df_events = events["train"].to_pandas()

# Load ELASTIC sync results
sync = load_dataset("luxury-lakehouse/obso-pausa-inputs", "elastic_sync")
df_sync = sync["train"].to_pandas()

# Join events to their best-matching tracking frame
merged = df_events.merge(df_sync, on=["match_id", "event_id"], how="inner")
print(f"{len(merged)} events with frame alignment")

Explore interactively: Soccer Analytics App

What Is This Dataset?

This dataset provides the input layer for OBSO and PAUSA computation. It contains two configs:

1. events — Match events extracted from IDSSE open data in DFL format (Play, KickOff, TacklingGame, etc.) with pitch coordinates in meters.
2. elastic_sync — The output of the ELASTIC algorithm (Kim et al. 2025) that synchronizes each event to its best-matching tracking frame, enabling event-tracking fusion without annotated event locations.

Together, these allow downstream pipelines to compute pitch control at the exact moment of each event by looking up the corresponding tracking frame.

Data Fields — events Config

Column	Type	Description
match_id	string	Match identifier
event_id	string	Unique event identifier
event_type	string	DFL event type (Play, KickOff, TacklingGame, etc.)
timestamp_seconds	double	Elapsed time in seconds from period start
period	int	Match period (1 or 2)
player_id	string	DFL PersonId of the acting player
team	string	Team affiliation (home or away)
x	double	Event x-coordinate (meters, 0–105)
y	double	Event y-coordinate (meters, 0–68)

Data Fields — elastic_sync Config

Column	Type	Description
match_id	string	Match identifier
event_id	string	Event identifier (join key to events config)
frame_id	int	Best-matching tracking frame number
alignment_confidence	double	ELASTIC alignment confidence (0–1, higher = more confident)
alignment_error_seconds	double	Estimated temporal error of the alignment in seconds

Coordinate System

Events: DFL pitch-origin meters (105×68m). The origin (0, 0) is at the center of one goal line; x runs along the pitch length (0–105m), y along the width (0–68m).

ELASTIC sync: Frame IDs reference 25fps IDSSE tracking data. Each frame_id maps to a specific instant in the tracking timeline.

Data Sources

Source	Matches	License
IDSSE Open Data	7	CC-BY 4.0

The IDSSE (Integrated Dataset of Spatiotemporal and Event Data in Elite Soccer) provides synchronized event and tracking data from German Bundesliga matches.

Companion Resources

Resource	Description
OBSO/PAUSA Values	Computed OBSO and PAUSA scores per pass
Pitch Control Tracking	Per-player per-frame pitch control from tracking data

Limitations

• Small sample: Only 7 Bundesliga matches from the IDSSE open data release. Results may not generalize across leagues or tactical systems.
• ELASTIC alignment accuracy: Synchronization quality depends on event timing precision in the source data. Events with ambiguous timestamps may have lower alignment_confidence.
• DFL event taxonomy: Event types follow DFL conventions, which differ from StatsBomb or Opta taxonomies. Cross-provider comparisons require mapping.
• No ball tracking in events: Event coordinates represent player position at the moment of action, not ball trajectory.

Citation

If you use this dataset, please cite the IDSSE and ELASTIC papers:

@article{bassek2025idsse,
  title={An integrated dataset of spatiotemporal and event data in elite soccer},
  author={Bassek, Manuel and Weber, Henrik and Rein, Robert and Memmert, Daniel},
  journal={Scientific Data},
  volume={12},
  pages={283},
  year={2025},
  publisher={Nature Publishing Group}
}

@inproceedings{kim2025elastic,
  title={ELASTIC: Event-Tracking Data Synchronization in Soccer Without Annotated Event Locations},
  author={Kim, Hyunsung and Theis, Fabian and Kim, Hanjun},
  booktitle={ECML-PKDD MLSA Workshop},
  year={2025},
  eprint={2508.09238},
  archiveprefix={arXiv}
}

More Information

Explore interactively: Soccer Analytics App

• License: MIT
• Publish script: scripts/publish_obso_pausa_inputs_hf.py