Feds Remove Crucial Prison Data Set, Leaving Researchers in the Dark

Researchers lost access to critical location data for carceral facilities throughout the country after the federal government suddenly discontinued it earlier this year.  

The Department of Homeland Security removed the Homeland Infrastructure Foundation-Level Data (HIFLD), which provided geospatial information about infrastructure throughout the country, from its website on August 26. It’s unclear whether the federal government will still maintain or update the dataset or who will still be able to access the data, but a message on the HIFLD website notes that any “DHS mission partner” can request ongoing access to “select Open layers” of the dataset.  

Researcher have relied on HIFLD data to study the geography of the American carceral system. Among other information contained in the dataset, it includes location data for prisons and jails throughout the nation. It also includes metadata recording the capacity of each carceral facility, the county the facility is in, and whether the facility is open or closed.  

The Behind Bars Data Project (BBDP) has relied on data from HIFLD to better understand carceral mortality. One current project our team is working on involves collecting data from the National Emergency Medical Services Information System (NEMSIS) about emergency calls to carceral facilities. NEMSIS data identifies facilities by ZIP code, so if a particular ZIP code has more than one facility, BBDP researchers can often use HIFLD data to determine which of the facilities placed the call.  

The data from HIFLD has some overlap with other data sets, including information about carceral facilities provided by the Bureau of Justice Statistics, but the HIFLD data is more complete and, until it was taken offline, was updated more frequently.  

The data had wide use for researchers studying carceral facilities and incarcerated populations. Nicholas Shapiro, an environmental researcher and associate professor at UCLA, worked on a project looking at harmful PFAS exposure at carceral facilities. The research, which showed that roughly one million incarcerated people (including 13,000 children) are likely exposed to dangerous PFAS chemicals, made international news.  

Shapiro used the HIFLD to find carceral facilities that lay downstream from known PFAS exposure sources, finding that roughly half of U.S. prisoners had some level of exposure to PFAS 

Shapiro also served as a data advisor for a project by The Intercept looking at the risks of climate change to people incarcerated in US facilities. Maps created by the news organization for the project relied on location data obtained through HIFLD. The data and mapping project formed the basis for a series of stories published by The Intercept about environmental harms in US carceral facilities, with headlines like “Boiling Behind Bars: In Sweltering Texas, Prisons Without Air Conditioning Are About to Get a Lot Hotter.” 

Shapiro said much of the research he’s done would have been impossible without the HIFLD dataset. The Intercept project, he said, would have been impratical if the HIFLD data was unavailable, since reporters would have had to map out the location of every carceral facility in the country.  

“Inherently, a lot of the data on carceral facilities across the country, the sort of archipelago of thousands of facilities… it’s usually siloed. So the data is siloed between city governments, county governments,” he said. “What’s unique about this data set is it has them all.”  

Many researchers, including Shapiro and the BBDP team, have access to archived copies of the latest version of the HIFLD data before it was taken offline. But as facilities open and close, the snapshotted data will gradually become outdated and, eventually, virtually useless. And newer researchers or members of the public hoping to study the carceral system may not know how or where to find the archived versions maintained by more experienced researchers.  

Shapiro said the removal of HIFLD data also creates an interoperability problem, as data from, for example, a particular state’s county jails won’t be formatted in the same way as the archived HIFLD data set, creating major workability issues when dealing with thousands of lines of data. Shapiro has worked on several projects integrating outside data, including data about ICE facilities, into the HIFLD dataset, and plans to soon launch a version of the dataset that includes Bureau of Indian Affairs jails.   

During the first Trump administration, Shapiro launched the Environmental Data and Governance Initiative, which invited researchers to archiving events where they downloaded and preserved data that they feared would be lost. Shapiro said the removal of the HIFLD data is of a piece with the administration’s degrading of data access in a variety of fields.  

“We’re also tracking the changes to government websites. We saw mentions of the Paris Climate Accord getting deleted,” he said. “We also saw the ways the federal government was not just deleting data but changing the interpretation of data. And it’s all part of a larger project to thwart accountability and oversight that unfortunately is very effective.”