Biography
My research examines the promise and limits of AI/ML-based tools to study sensitive topics of substantial interest to the public. I study policing, specifically how two-way radio use relates to officer coordination and behavior. My work seeks to clarify what happens before and during officer encounters with the public by converting unstructured radio transmission data into structured data for study of how policing operates in practice. My goal is to help others make sense of police use of radio as an understudied sociotechnical system critical to policing.
Since little work has explored how officers use two-way radio in practice, much of my work involves navigating the notable lack of legal guidelines covering publicly available data which, nonetheless, contains sensitive information about members of the public.
My aspirations include (a) development of decentralized participatory data governance policies, with initial focus on policing data; (b) deployment of explainable models for processing unstructured policing data, specifically radio transmissions; (c) engagement with community members to determine standards for responsible access to policing data and identify research questions that address the needs and concerns of local residents.
I am currently on the tenure-track faculty job market, but I also love to do time-bounded projects!
- Computational Social Science
- Translational AI
- AI Ethics
- Urban Sociology
- Social Networks
- Spatial Analysis
PhD in Sociology, 2016
University of Chicago
MA in Social Science, 2005
University of Chicago
BS in Mathematics, 2004
Virginia Tech
BS in Physics, 2004
Virginia Tech
Experience
Responsibilities include:
- Direct and oversee development of voice activity detection (VAD), automatic speech recognition (ASR), speech emotion recognition (SER), and natural language processing (NLP) strategies to determine the viability of using publicly available broadcast police communications to understand police behavior via Spencer’s phenomenological variant of ecological systems theory (PVEST).
- Direct all aspects of manuscript development, from ideation to submission, across relevant fields and publication venues.
- Mentor 15-20 (graduate) research assistants engaged in qualitative, quantitative, and computational tasks supporting theory-based data science
Responsibilities included:
All aspects of quantitative and mixed methods research projects, including but not limited to production of IRB protocols, secondary data analysis, and experiment design, in service of the Urban Resiliency Initiative and in collaboration with on-site and off-site research teams.
- Supported development of PVEST-informed mixed methods based research designs
- Developed a high-level PVEST-based research strategy for the NSF-funded STEM-US Center
- Assessed efficacy of an arts-augmented STEM curriculum on academic performance of 7th-8th graders
- Evaluated assessment tool to identify the needs of HBCU STEM students in order to improve success of HBCU students in STEM careers
- Contributed to multiple successfully funded grants to conduct transdisciplinary research between education researchers, social scientists, computer scientists, and physical scientists
Objective: Geocode full count census data from 1900 to 1940 for 69 cities
- Calculate residential segregation measures via point, line, and polygon data
- Assess state-of-the-art segregation measures given newly availabile point data
- Design and implement ETL workflow using Python
- Integrate legacy Python and R code into workflow
- Develop metrics to evaluate data integrity/quality
- Optimize process for scaling to 69 cities in 1930 (>30M records)
- Use Python, Stata, and Matlab to…
- Validate accuracy of digitized street names using additional data
- Perform fuzzy string matching to correct street names using additional data
- Extract Census block numbers from historic maps using machine vision
- Report relevant metrics to assess process performance
- Analyze residential segregation patterns in 1930 St. Louis for 10+ racial/ethnic groups