Research

Open Datasets

Structured, downloadable, CC BY 4.0. Take them, check them, build on them. We would rather be corrected than cited uncritically.

Licence and citation#

Both datasets are released under Creative Commons Attribution 4.0. Use them commercially, modify them, redistribute them. The only condition is attribution.

Cite as: ProBirdFeeder Feeder Bird Trait Matrix, probirdfeeder.com/research/datasets/feeder-bird-traits/, accessed [date]. For the observation network: ProBirdFeeder Feeder Log, probirdfeeder.com/research/feeder-log/, accessed [date], noting the row count at the time you pulled it, because it changes.

Why we publish the build script too#

A dataset you cannot regenerate is a dataset you have to trust. Ours ships with research/build_dataset.py, which reads the published species profiles and emits the CSV deterministically. Run it against this site and you get byte-identical output, or you find a bug, which is equally useful to us. The extraction rules it implements are documented in plain English on the methods page.

Formats#

FileFormatWhat it is
feeder-bird-traits.csvCSVOne row per species, list columns pipe-delimited
feeder-bird-traits.jsonJSONSame data with lists as arrays, plus the source text each tag came from
matrix-food.jsonJSONSpecies slug to food tags, for joins
matrix-feeder.jsonJSONSpecies slug to feeder-type tags, for joins
summary.jsonJSONPrecomputed aggregates used by the findings pages
feeder-log.csvCSVLive export of the observation network
api/feeder-log/summary.jsonJSONLive aggregate counts, CORS enabled