Datasets

We collect datasets from various sources and prepare them taking into account a potentially discriminated group, and an attribute/score in which discrimination can be embodied. Click on the name of a dataset below for details and availability. Some datasets are directly available for download, while others are available upon requests for privacy reasons.

Dataset	Attribute/Score	# Entries	Format
COMPAS Recidivism Risk	recidivism	18K	csv
Statlog (German Credit -SCHUFA)	credit rating	1k	csv
SAT	test score	1,6 k	csv, pdf
PSU-Chile	test score	2,9 M	csv
Adult Census	income	48,8 k	csv
SOEP	multiple	depending on wave	multiple

If you have a dataset that you would like to contribute, please get in touch with us.