Meteorology

Leuven.cool Citizen Science Urban Weather Station Network

Over 1 billion raw observations from ~110 low-cost weather stations spread across Leuven (Belgium), at 16-second resolution from June 2019 to March 2025.

Size
12.9 GB
Format
CSV, CSV.gz
License
CC-BY-NC-4.0
Access
Open Access
Source
Zenodo (record 14893734)

License note: Non-commercial

Your edits to the notebook only affect your own copy in the Hub. Changed it and want the original back? Restore it here.

Challenge

Is there a measurable urban heat-island effect across Leuven's citizen weather network - are some stations systematically warmer than others, and does that gap open up at night when built-up ground releases the day's stored heat?

Where to start
Do NOT load a raw file whole (each quarter is hundreds of MB gzipped, billions of rows in total). Load the small STATIONS.csv for coordinates (ID, LATITUDE, LONGITUDE, ALTITUDE), then use DuckDB's read_csv_auto to aggregate ONE summer quarter (RAWDATA2019Q3.csv.gz) straight off disk: convert TEMPF to Celsius with (TEMPF-32)*5/9, filter implausible values, shift DATEUTC by +2h to local CEST, and GROUP BY station ID and hour-of-day into a ~110-station x 24-hour temperature grid.
What to share
A station coordinate map coloured by night-time temperature anomaly, plus a plot of the between-station spread (std of station means) versus hour of day showing whether the gap peaks in the small hours. Report the warmest-minus-coolest night gap and name which DuckDB/chunking technique you used to stay inside the 4 GB budget; optionally repeat on a winter quarter (RAWDATA2020Q1.csv.gz) to test whether the heat island is stronger in summer.

💬 Discuss this dataset, ask questions and share your results in its discussion thread. Challenges are open-ended — there's no single right answer.