GoldenCheetah

Data from the GoldenCheetah OpenData Project in Parquet format.

Table	Contents
measurements	Source CSV filename, time (s), distance (km), power (W), heart rate (bpm), cadence (rpm), altitude (m)
activities	Original activity fields, with METRICS and XDATA stored as JSON
athletes	Athlete ID, gender, birth year, and source version

athlete_id identifies the athlete in each table. source_file identifies each CSV trace. Activity dates and CSV filenames use the source timestamps.

Use

Download the Python scripts and requirements.txt, then install:

python3 -m venv .venv
.venv/bin/python -m pip install -r requirements.txt

Convert ZIP files:

.venv/bin/python prepare_dataset.py /path/to/zips/*.zip --output data

Query the public data:

.venv/bin/python query_dataset.py --remote --sql "SELECT sport, COUNT(*) FROM activities GROUP BY sport"

Query a source metric:

SELECT date, TRY_CAST(METRICS::JSON ->> '$.average_power[0]' AS DOUBLE) AS power
FROM activities;