Facility Data & Crosswalks

Background

The UCLA Law Covid-19 Behind Bars Data Project collects and reports facility-level data on COVID-19 in prisons, jails, and other carceral facilities. Our core dataset can be found on GitHub here. This repository contains the facility datasets and crosswalks used to integrate data from various sources.

There are very few national datasets on correctional facilities in the U.S. The lack of quality and contemporary descriptive data for correctional facilities poses a major problem for research on corrections, particularly research tracking the spread of Covid in these facilities.

The UCLA team developed these datasets to uniformly describe the entities reported by various government agencies by combining information from various sources. This includes information from national datasets, like the Homeland Infrastructure Foundation-Level Data (HIFLD) Prison Boundaries dataset produced by the Department of Homeland Security and datasets produced by the Bureau of Justice Statistics. We have also filled in gaps in these sources by independently gathering data through public records requests and manual investigation.

Note: These files continue to evolve as reporting agencies post new entities and new spellings of those entities. As such, the UCLA team continuously updates these datasets.

Directory Structure

This repository contains two datasets managed by the UCLA Law Covid-19 Behind Bars Data Project:

Facility Information Dataset

• Data File: data/fac_data.csv
• Description: This dataset contains detailed descriptive information on each entity in our dataset (e.g. type of entity, capacity, population, geographic information, etc.). Each row represents a unique entity.

Facility Spellings Crosswalk

• Data File: data/fac_spellings.csv
• Description: This crosswalk maps alternative spellings of a given entity to the cleaned name in the Facility Information Dataset. Each row represents a unique spelling of a given entity, as reported by an agency that we collect data from.

Data Dictionary

Note that we use facility and entity interchangeably here. While the majority of the rows in these datasets are facilities (e.g. jails or prisons), several rows are geographic entities because an agency reports data at that level.

Facility Information Dataset

Variable	Description
Facility.ID	Integer ID that uniquely identifies every facility
State	State where the facility is located
Name	Cleaned name for the facility
Jurisdiction	Jurisdiction of the reporting agency Potential values: state, county, federal, immigration
Description	Entity type, designated by UCLA staff (see below for more detailed descriptions) Potential values: Geographic, Administrative, Prison, Jail, Hybrid, Reception Center, Transitional Center, Medical Facility, Detention Center, Prison Unit, Work Camp, Aged and Infirmed, Contractor
Security	Security level of the facility, designated by UCLA staff Potential values: Max, Med, Min, Max/Med, Max/Min, Med/Min
Age	Age group kept in the entity if known, designated by UCLA staff Potential values: Adult, Juvenile, Mixed
Gender	Gender group kept in the entity if known, designated by UCLA staff Potential values: Female, Male, Mixed
Is.Different.Operator	Binary indicator for whether the entity is run by an organization different from the reporting jurisdiction (e.g. a private company or a county government)
Different.Operator	Name of the organization operating the entity if run by an organization different from the reporting jurisdiction
Population.Feb20	Population of the facility as close to February 1, 2020 as possible. This variable is a combination of HIFLD values and data gathered by UCLA staff through public records requests. Population values gathered by UCLA staff were prioritized over data from HIFLD if both were available.
Capacity	Capacity of the facility if known. This variable is a combination of HIFLD values and data gathered by UCLA staff. Capacity values gathered by UCLA staff were prioritized over data from HIFLD if both were available.
HIFLD.ID	Facility's corresponding Homeland Infrastructure Foundation-Level Data ID
BJS.ID	Facility's corresponding Bureau of Justice Statistics ID
Source.Population.Feb20	Population source Potential values: HIFLD, Public Records
Source.Capacity	Capacity source Potential values: HIFLD, Public Records
Address	Facility's address
City	Facility's city
Zipcode	Facility's zipcode
Latitude	Facility's latitude
Longitude	Facility's longitude
County	Facility's county
County.FIPS	Facility's 5-digit county FIPS code
Website	Facility's website. This corresponds to either the website where COVID-19 data is being reported, or the general website maintained by the facility.
ICE.Field.Office	3 letter code for ICE enforcement and removal office only shuld be present if Jurisdiction is immigration example of codes here

Facility Spellings Crosswalk

Variable	Description
Facility.ID	Integer ID that uniquely identifies every facility
xwalk_name_raw	Alternative spelling for the facility
xwalk_name_clean	Cleaned name for the facility
State	State where the facility is located
Jurisdiction	Jurisdiction of the reporting agency Potential values: state, county, federal, immigration
Source	Source where the alternative spelling appears

Entity Types

A key variable in the Facility Information Dataset is the entity type designation (Description). Because agencies report data at varying units of aggregation, this variable provides meaningful categorizations for the entities we collect data on. More detailed descriptions of each designation is provided below.

Designation	Entity Description
Geographic	Cover a geographic region and are not a particular facility (e.g. statewide values, parole regions).
Administrative	Staff specific and do not hold correctional residents (e.g. agency headquarters, training academies).
Prison	Generally hold individuals who have been convicted of a crime and sentenced. These are exclusively run by state or federal agencies.
Jail	Generally hold individuals who have not been convicted of a crime and are detained or awaiting trial. These facilities also house individuals who have been convicted of a crime but are serving a short sentence. These are generally run by county governments.
Hybrid	Generally hold both individuals convicted of a crime and individuals who have not been convicted of a crime. These are facilities housed in jurisdictions that combine their prison and jail systems, like Hawaii and DC.
Reception Center	Prisons which also serve as intake centers for newly convicted individuals. At these facilities, individuals are designated a security level and assigned to a long-term facility. These facilities can also hold their own long-term populations.
Transitional Center	Prisons which also serve as re-entry centers for convicted individuals who will be released. These facilities often have transitional programs where individuals are technically trained or work in local communities through occupational programs.
Medical Facility	Prisons which have significant inpatient and outpatient care services for convicted individuals and typically are correctional hospitals.
Detention Center	Generally hold individuals who have not been convicted of a crime and are detained or awaiting trial or deportation hearings. These are generally run by state or federal agencies.
Prison Unit	Units within prisons.
Work Camp	Prisons that house individuals who generally work on significant labor projects like farming, land management, and firefighting.
Aged and Infirmed	Prisons that generally house older individuals and individuals with significant chronic diseases.
Contractor	Non-government agencies (e.g. private companies, charities, etc.) that report data to a correctional agency. Many of these agencies provide services like drug treatment, job placement, and halfway housing among others.

Update Instructions

These instructions are intended to be internal reference for the UCLA project team. To update the facility crosswalks with a new facility or new alternative spelling:

1. Add the new entries to the Google Sheet here, following the instructions in the first tab
2. Run R/main.R, which does the following:

• Reads from the Google Sheet
• Populates missing data if possible (e.g. from the HIFLD database)
• Performs validation checks on the new entries
• Assigns a Facility.ID to new facilities
• Updates the facility crosswalk csv files in the data folder

You should run main.R in the order specified (i.e. update fac_data before updating fac_spellings) when adding both facilities and spellings at the same time.

Note that main.R assumes that you have the following:

• Access to the Google Sheet, which will automatically generate a valid OAuth token for googlesheets4
• A geocodio API key, available from here
• An updated version of the behindbarstools R package

Citations

Citations for academic publications and research reports:

Sharon Dolovich, Aaron Littman, Kalind Parish, Grace DiLaura, Chase Hommeyer, Michael Everett, Hope Johnson, Neal Marquez, and Erika Tyagi. UCLA Law Covid-19 Behind Bars Data Project: Jail/Prison Confirmed Cases Dataset [date you downloaded the data]. UCLA Law, 2020, https://uclacovidbehindbars.org/.

Citations for media outlets, policy briefs, and online resources:

UCLA Law Covid-19 Behind Bars Data Project, https://uclacovidbehindbars.org/.

License

Our data is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. That means that you must give appropriate credit, provide a link to the license, and indicate if changes were made. You may not use our work for commercial purposes, which means anything primarily intended for or directed toward commercial advantage or monetary compensation.