Asylum Case Predictor

Home | About

In the United States, foreign nationals who fear persecution in their home country can apply for asylum under the Refugee Act of 1980.

Data Source

The project uses a large data set secured under a Freedom of Information Act (FOIA) request detailing United States Asylum court decisions going back multiple decades and machine learning to predict the outcomes of the cases.

Prediction Accuracy

We are able to predict the final outcome of a case with 80% accuracy at the time the case opens. Using data available on the decision date, our model correctly classifies 82% of all refugee cases by 2019.

This app should be used with the caveat that it renders only a prediction based on the inputs you provide.

Supporting Research

Early Predictability of Asylum Court Decisions

Can Machine Learning Help Predict the Outcome of Asylum Adjudications?

Related Links

Wikipedia: Refugee roulette

Contributors

Matthew Thomas Dunn
Elliot Ash
Zhang Sai
Matthew Mauer
Madeline Bassetti

Special Thanks

Jaya Ramji-Nogales

A collaboration by: Daniel Chen, Lev Tatz, etc