Facial Profiling: Defendant Physical Characteristics, Machine Learning Analytics, and Criminal Justice Disparities in Miami-Dade County


This project seeks to document the extent to which colorism shapes criminal justice outcomes and facial recognition technologies by analyzing the mugshots of 200,000 arrestees linked to court records in Miami-Dade County. In working with the Miami-Dade Public Defender’s Office, if the team finds that skin color discrimination in Miami courts exists, it could constitute a credible enough showing of discrimination to spur litigation. Also, since the Miami-Dade County Public Defender is committed to evidence-based methods for reducing disparities in the county’s criminal justice system, the team’s results could spur equal protection litigation surrounding colorism in courts or the use of facial recognition technology. In addition, the team hopes to encourage local police to adopt evidence-based guidelines for the use of facial recognition technology.

 Phase I (2019)

Race-based disparities in criminal justice outcomes are a pervasive problem in the United States. Despite increased societal concern about racial and ethnic disparities, there is a disturbing lack of research on the topic. Combining Computer Science with Sociology and Law, the U-LINK ‘Race and Facial Profiling’ team will determine how defendant physical characteristics and facial recognition software influence criminal justice outcomes. Using an unprecedented data set combining mugshot photos and court records in Miami-Dade County, team members will work together to develop a machine learning model to test whether arrestees’ physical features lead to disparate punishment outcomes for racial/ethnic minorities. This work will also shed light on how facial recognition software “sees” race/ethnicity, of critical importance as more and more police agencies use this software. Ultimately the team hopes to substantively advance racial and ethnic justice; “If we can show that racial disparities exist in a sample of our size,” the team writes in their proposal, “…[it] might bring to light improper race-based decision-making, which could then be challenged in the courts.”


Nicholas Petersen, Sociology; Tamara Lave, School of Law; Ubbo Visser, Computer Science