Nason/cook-sao-database

Cook County State's Attorney Felony Case Database

The Cook County (Chicago) State's Attorney's Office published five case-level datasets covering every felony case it received from 2011 until it closed the series on 2024-12-30: intake (felony review), initiation (charges, arraignment, bond), dispositions, sentencing and diversion referrals. Each is one table here, typed from the portal's own column metadata and joinable on case_id / case_participant_id / charge_id. 3,171,690 rows across 5 tables.

Table	Description	Row Count	Column Count
initiation	One row per charge on a case at initiation (after felony review approval), 2011-2024: charge statute/class, initiation event and date, arraignment date, initial and current bond type/amount, demographics	1,228,260	38
dispositions	One row per charge disposition, 2011-2024: charged offense as disposed, disposition and reason, judge, court, plus the case-level intake fields	1,080,014	33
intake	One row per potential defendant brought to the State's Attorney for felony review, 2011-2024: received date, offense category, felony-review result, arrest and incident dates, demographics, arresting agency	528,111	17
sentencing	One row per sentence imposed on a disposed charge, 2011-2024: sentence type, commitment type/term/unit, sentence judge and court, whether it is the current sentence, case length	305,884	41
diversion	One row per diversion-program referral, 2011-2024: program, referral date, primary charge, result, closed date	29,421	13

Query it remotely

INSTALL httpfs; LOAD httpfs;
ATTACH 'https://huggingface.co/datasets/Nason/cook-sao-database/resolve/main/cook-sao.duckdb' AS cook_sao (READ_ONLY);

SELECT EXTRACT(YEAR FROM received_date) AS year, felony_review_result, COUNT(*) AS n
FROM cook_sao.intake GROUP BY 1, 2 ORDER BY 1, 3 DESC;

R with DBI and duckdb

Install the dependencies once with install.packages(c("DBI", "duckdb")).

library(DBI)
con <- dbConnect(duckdb::duckdb())
dbExecute(con, "INSTALL httpfs")
dbExecute(con, "LOAD httpfs")
dbExecute(con, "ATTACH 'https://huggingface.co/datasets/Nason/cook-sao-database/resolve/main/cook-sao.duckdb' AS cook_sao (READ_ONLY)")
dbGetQuery(con, "SELECT received_date, felony_review_result FROM cook_sao.intake LIMIT 5")
dbDisconnect(con, shutdown = TRUE)

DuckDB uses HTTP range requests; queries run on your machine and may need substantial memory or download data. A LIMIT does not guarantee a small scan.

Or with the datapond packages: pip install datapond / pak::pak("datapond-db/datapond-r").

Build pipeline, the SAO data glossary and caveats: https://github.com/ian-nason/cook-sao-database Source: Cook County open-data portal, https://datacatalog.cookcountyil.gov (public domain).