Research

Graduate research in crime, terror, and intelligence.

Selected work from the research portfolio I compiled as my capstone for my M.S. in Criminal Justice at Boston University. It ranges from statistical and geospatial analysis to policy proposals, each built toward recommendations someone could act on.

329 homicides mapped across three LAPD divisions
Boston University · M.S. Criminal Justice · 2021, revised 2026

Violence Reduction Initiative

A COMPSTAT-style briefing on homicide and firearm violence in the LAPD's 77th Street, Southwest, and Newton divisions from 2016 to mid-2019. I mapped where, when, and how the violence happened, narrowed it to specific hot spots and reporting districts, and recommended targeted responses.

Data
Public LAPD crime data and calls for service, January 2016–June 2019
Methods & tools
Hot spot mapping, day and shift patterns, and trend comparisons, analyzed in SPSS and ArcGIS and visualized in Power BI
Concentration
77th Street accounted for 49% of homicides; 77% of recent homicides involved a firearm
Trend
Firearm incidents rose 90% from February to May 2019, against a 52% average for those months
Hot spots
Vermont & Florence, McKinley Ave, Hillcrest & MLK, and reporting district 1243 near Mann Middle School
Recommendations
More patrols in Hyde Park, a community watch around Mann Middle School, and cameras and lighting at Vermont & Florence
  • COMPSTAT
  • Hot spot analysis
  • Geospatial mapping
  • Trend analysis
  • SPSS
  • ArcGIS
  • Power BI
Read the full brief Download PDF · 12 pages

A simulated briefing built on public LAPD data, not an official LAPD product.

17,965 use-of-force incidents from 13 U.S. agencies
Boston University · M.S. Criminal Justice · 2021, revised 2026

Police Use of Force in the United States: Non-Lethal Use of Force Analysis

What predicts whether an officer reaches for a Taser or pepper spray? With two classmates, I tested eight officer, suspect, and call characteristics across 17,965 use-of-force incidents, first one at a time and then together in logistic regression models.

Data
17,965 incidents from 13 U.S. law enforcement agencies, 1992–2007 (Taylor & Koper, 2013)
Methods & tools
Chi-square tests of independence and binary logistic regression, run in SPSS
Strongest predictor
Suspect behavior: verbal resistance and mild aggression raised the odds of pepper spray use about 10 times, and fleeing raised the odds of Taser use nearly 5 times
Race
After controls, Black suspects had higher odds of pepper spray use and Hispanic suspects higher odds of Taser use than White suspects
Not significant
Officer experience and suspect age, once the other factors were controlled
Limitations
Up to 65% missing data on some variables, which the report addresses directly
  • Logistic regression
  • Chi-square tests
  • Descriptive statistics
  • SPSS
  • Use-of-force policy
Read the full report Download PDF · 41 pages

Co-authored with Cuvee Zhou and Kelly Feeney.

More research

Want to talk through the research?

Connect on LinkedIn Back to home