Immigration and Customs Enforcement

ICE collects data on every person it encounters, arrests, detains, transports via flight, and deports, and every detainer request to local jails.

We post lightly processed versions of the arrests, detainer requests, and detentions datasets to facilitate analysis. We also post a more heavily processed version of the detention data, collapsed to the detention stay level, to make analysis more straightforward.

Getting started with ICE data

Our ICE frequently asked data questions is a good place to start before analyzing the data, and this page describes the steps we have taken to process the datasets on this page. We also provide a codebook that describes each data table and the fields within it.

We recommend citing these data as “government data provided by ICE in response to a FOIA request, processed by the Deportation Data Project, and analyzed by [your organization].”

Latest data release

The most recent release of data covers ICE enforcement actions through early March 2026. We are actively seeking ongoing updates.

Explore the data in our interactive tool

We recommend using the detention stays dataset for most detention analyses at the individual level. For analyses at the detention facility level, by contrast, we recommend using the detention stints dataset. We do not provide an Explore tool for the detention stints dataset because filtering by stint might lead to inaccurate individual-level inferences.

Code and notes on how we processed the data are available on the ICE processed data documentation page.

We have several concerns about this new dataset.

First, and most important, we continue to have questions about the reliability of the encounters and removals data. These encounters and removals datasets are similar to the October versions but different from the late July release that we believe was most reliable. A comparison of the October and late July removals datasets is available here.

Second, in this version of the dataset, ICE omitted detention stints in hospitals and medical centers. This affects a small number of stints relative to the total—and therefore should not prevent analysis of the detention data—but it makes analysis of hospital stints impossible.

Third, ICE has redacted the “case_category” values for some individuals; comparing this release to previous releases suggests that these are mostly individuals who had credible fear interviews. For some of the individuals, ICE has also redacted detention release reasons (in the detentions table) and processing disposition values (in the encounters and removals tables). To obtain some of this missing information, we suggest joining the detention and removals tables to determine which individuals were booked out to be removed.