Datasets:

mstz
/

compas

is_male bool	age int64	race string	number_of_juvenile_fellonies int64	decile_score int64	number_of_juvenile_misdemeanors int64	number_of_other_juvenile_offenses int64	number_of_prior_offenses int64	days_before_screening_arrest int64	is_recidivous bool	days_in_custody int64	is_violent_recidivous bool	violence_decile_score int64	two_year_recidivous int64
true	69	Other	0	1	0	0	0	-1	false	7	false	1	0
true	34	African-American	0	3	0	0	0	-1	true	10	true	1	1
true	44	Other	0	1	0	0	0	0	false	1	false	1	0
true	43	Other	0	4	0	0	3	-1	false	12	false	3	0
false	39	Caucasian	0	1	0	0	0	-1	false	3	false	1	0
true	21	Caucasian	0	3	0	0	1	428	true	1	true	5	1
true	27	Caucasian	0	4	0	0	0	-1	false	1	false	4	0
false	37	Caucasian	0	1	0	0	0	0	false	1	false	1	0
true	41	African-American	0	4	0	0	0	-1	false	1	false	2	0
true	37	Hispanic	0	1	0	0	0	0	false	0	false	1	0
true	25	African-American	0	10	0	0	3	-1	false	62	false	9	0
true	31	Caucasian	0	5	0	0	6	-1	true	1	true	4	1
false	21	Caucasian	0	4	0	0	0	-2	false	1	false	5	0
true	43	Caucasian	0	1	0	0	1	-1	false	1	false	2	0
true	32	Other	0	3	0	0	0	-1	false	2	false	4	0
true	27	African-American	0	3	0	0	8	-1	true	1	true	3	1
false	36	Caucasian	0	3	0	0	3	53	false	1	false	1	0
true	33	African-American	0	10	0	0	0	0	true	21	true	6	1
true	55	Caucasian	0	1	0	0	0	-1	false	2	false	1	0
true	26	Caucasian	0	8	0	2	6	81	false	12	false	8	0
true	29	Caucasian	0	2	0	0	0	0	false	0	false	2	0
false	21	African-American	0	8	0	0	2	-1	false	1	false	8	0
true	51	African-American	0	1	0	0	2	-1	false	1	false	1	0
false	51	African-American	0	2	0	0	7	-1	false	8	false	2	0
true	25	Caucasian	0	10	0	0	9	-1	true	9	true	9	1
true	35	African-American	1	8	0	4	13	-1	true	1	true	3	1
true	49	Caucasian	0	1	0	0	0	0	false	0	false	1	0
true	29	African-American	0	4	0	0	0	0	false	0	false	3	0
false	28	Other	0	2	0	0	0	-1	false	1	false	2	0
false	63	Hispanic	0	1	0	0	1	-219	false	8	false	1	0
true	31	African-American	0	5	0	1	15	-1	true	136	false	7	0
true	30	Hispanic	0	7	0	0	0	0	false	23	false	3	0
true	49	African-American	0	1	0	0	4	-1	false	1	false	1	0
true	53	Caucasian	0	5	0	0	8	-13	false	13	false	2	0
true	35	African-American	0	3	0	0	5	-1	false	1	false	2	0
true	62	Hispanic	0	1	0	0	0	0	false	0	false	1	0
true	33	Other	0	1	0	0	0	-1	false	1	false	1	0
false	56	African-American	0	1	0	0	0	0	false	3	false	1	0
true	45	Hispanic	0	1	0	0	6	-1	true	242	true	1	1
true	22	Hispanic	0	4	0	0	1	-1	false	1	false	6	0
true	39	African-American	0	4	0	0	1	-1	false	5	false	2	0
true	40	African-American	0	2	0	0	2	-15	false	3	false	1	0
true	22	Caucasian	0	8	0	0	0	-1	false	21	false	6	0
false	26	African-American	0	4	0	0	1	-1	true	58	true	3	1
true	27	Caucasian	0	7	0	0	1	-43	true	137	true	8	1
true	32	African-American	0	2	0	0	0	-1	false	1	false	1	0
true	41	African-American	0	2	0	0	1	-1	false	2	false	2	0
true	30	African-American	0	10	0	0	8	-1	true	1	true	9	1
true	25	African-American	1	10	6	1	14	-1	true	19	true	10	1
true	20	Other	0	5	0	0	0	-1	true	423	true	8	1
false	35	African-American	0	2	0	0	1	-1	false	1	false	1	0
true	25	Caucasian	0	8	0	0	9	-1	true	2	true	9	1
false	20	Caucasian	0	7	0	0	0	-1	false	1	false	8	0
true	66	Caucasian	0	1	0	0	3	-64	false	30	false	1	0
false	54	Caucasian	0	1	0	0	0	0	false	0	false	1	0
false	44	Hispanic	0	1	0	0	0	-2	false	1	false	1	0
true	19	African-American	0	4	0	0	1	-10	true	6	true	7	1
true	24	African-American	0	2	0	0	1	-137	false	0	false	3	0
true	30	African-American	0	2	0	0	1	-1	false	1	false	2	0
false	39	Caucasian	0	1	0	0	0	-1	true	2	true	1	1
true	30	Caucasian	0	4	0	0	0	-1	false	27	false	2	0
true	31	Caucasian	0	5	0	0	0	-1	false	39	false	2	0
false	39	Caucasian	0	4	0	0	1	-1	false	5	false	3	0
true	59	African-American	0	9	0	0	14	178	true	110	true	4	1
true	43	Caucasian	0	1	0	0	1	-30	false	1	false	1	0
true	34	Caucasian	0	5	0	0	0	0	false	69	false	5	0
true	37	African-American	0	2	0	0	3	0	false	7	false	1	0
true	31	African-American	0	7	0	0	1	-1	false	1	false	8	0
false	30	Other	0	2	0	0	1	-1	false	8	false	2	0
true	31	Caucasian	0	10	0	0	13	-1	true	471	true	7	1
true	56	African-American	0	9	0	0	8	67	false	80	false	3	0
true	52	Caucasian	0	9	0	0	3	-2	false	1	false	6	0
false	64	Caucasian	0	1	0	0	2	-47	false	0	false	1	0
true	42	Caucasian	0	3	0	0	6	-4	false	86	false	1	0
false	25	Caucasian	0	4	0	0	2	-1	false	21	false	5	0
true	30	Caucasian	0	2	0	0	1	-20	false	4	false	2	0
false	37	Other	0	2	0	0	2	-1	false	2	false	1	0
false	24	African-American	0	2	0	0	0	-1	false	2	false	3	0
false	30	Hispanic	0	5	0	0	4	-35	true	241	true	5	1
true	31	African-American	0	2	0	0	1	-1	true	1	false	3	0
false	20	African-American	1	10	0	0	1	0	false	9	false	10	0
false	43	Caucasian	0	1	0	0	0	0	false	1	false	1	0
true	32	African-American	0	6	0	0	11	-2	true	36	true	5	1
true	33	African-American	0	5	0	0	8	-1	true	0	true	2	1
false	20	African-American	0	6	0	0	0	0	false	0	false	8	0
false	26	Caucasian	0	6	0	0	0	-2	false	1	false	3	0
true	63	Hispanic	0	1	0	0	1	-26	false	15	false	1	0
false	32	African-American	0	10	0	0	4	-7	true	118	true	9	1
true	29	Hispanic	0	4	0	0	1	-1	false	1	false	4	0
true	55	African-American	0	1	0	0	3	-51	false	1	false	1	0
true	39	Caucasian	1	4	0	0	1	-1	false	4	false	4	0
true	30	African-American	0	7	0	0	1	0	false	3	false	8	0
true	48	Caucasian	0	2	0	0	2	-1	false	1	false	1	0
true	22	African-American	0	2	0	0	1	-1	false	188	false	4	0
false	49	Caucasian	0	6	0	0	6	-56	false	5	false	2	0
true	31	Caucasian	0	5	0	0	0	0	false	50	false	4	0
true	23	Hispanic	0	3	0	0	0	-1	false	0	false	5	0
false	32	African-American	0	7	0	0	5	-1	false	1	false	3	0
true	52	Hispanic	0	1	0	0	2	-1	false	35	false	1	0
true	62	African-American	0	4	0	0	8	0	false	16	false	1	0

End of preview. Expand in Data Studio

Compas

The Compas dataset for recidivism prediction. Dataset known to have racial bias issues, check this Propublica article on the topic.

Configurations and tasks

Configuration	Task	Description
encoding		Encoding dictionary showing original values of encoded features.
two-years-recidividity	Binary classification	Will the defendant be a violent recidivist?
two-years-recidividity-no-race	Binary classification	As above, but the `race` feature is removed.
priors-prediction	Regression	How many prior crimes has the defendant committed?
priors-prediction-no-race	Binary classification	As above, but the `race` feature is removed.
race	Multiclass classification	What is the `race` of the defendant?

Usage

from datasets import load_dataset

dataset = load_dataset("mstz/compas", "two-years-recidividity")["train"]

Features

Feature	Type	Description
`sex`	`int64`
`age`	`int64`
`race`	`int64`
`number_of_juvenile_fellonies`	`int64`
`decile_score`	`int64`	Criminality score
`number_of_juvenile_misdemeanors`	`int64`
`number_of_other_juvenile_offenses`	`int64`
`number_of_prior_offenses`	`int64`
`days_before_screening_arrest`	`int64`
`is_recidivous`	`int64`
`days_in_custody`	`int64`	Days spent in custody
`is_violent_recidivous`	`int64`
`violence_decile_score`	`int64`	Criminality score for violent crimes
`two_years_recidivous`	`int64`

Downloads last month: 56

Size of downloaded dataset files:

831 kB

Size of the auto-converted Parquet files:

223 kB

Number of rows:

18,136