EOIR data release: simplified, open-source removal case data
We have released a simplified, open-source version of the CASE dataset, which the Executive Office for Immigration Review (EOIR) posts monthly. Although EOIR has posted the dataset monthly since 2018, it has been difficult to use. The version posted by EOIR is a zipped folder containing over 30GB of data spread across 18 large, linkable CSV files with limited documentation and with formatting errors that make opening some of the files challenging. Once opened, analyzing the data requires technical and substantive choices about which tables to use and how to link them. As a result of these difficulties, the data has mostly been used indirectly, through visualization tools, tabulation tools, or nontransparently preprocessed versions.
We address these issues by transparently creating a single dataset that includes one row for every removal case started in immigration court since fiscal year 1998. In addition to information on merits outcomes before immigration judges, we also include information on bond outcomes, applications for relief, immigration charges, appeals, and detention. The simplified data may be useful to journalists, advocates, and researchers.
The simplest way to start is with our online filtering tool (demonstrated in the video below), which allows filtering by variables such as nationality, court location, case outcome, and date before downloading a subset in Excel, Parquet, or Stata format without needing to download or process the full dataset. Advanced users can reproduce every step of our processing pipeline via our GitHub archive; every decision is documented and replicable from our posted code. We aim to update the dataset promptly after EOIR does so each month.
Demo of Explore data tool for the simplified EOIR case data.