Socioeconomic Drivers of Adult Disproportionate Minority Contact; Data Integrity in Algorithmic Policing 15 views

The intersection of engineering and social justice requires a fundamental shift in how data is perceived, not as a neutral mirror of reality, but as a byproduct of human and institutional history. This synthesis connects a technical investigation of Disproportionate Minority Contact (DMC) in the Charlottesville-Albemarle region with an STS critique of the "dirty data" that fuels modern predictive policing. Together, these projects demonstrate that engineering practice is inherently sociopolitical; the choice of variables in a model is not merely a technical decision but an ethical one that determines whether an algorithm serves to protect the public or automate the banishment of its most vulnerable members. By integrating geospatial engineering with STS frameworks, I argue that engineers must move beyond "black-box" efficiency and instead prioritize data integrity and social safety net guardrails to fulfill their professional duty to the public welfare. In my STS research, I investigated the ethical and systemic failures of predictive mapping technologies, specifically focusing on how historical arrest records, often referred to as "dirty data," perpetuate cycles of biased policing. My analysis revealed that many current algorithms operate on the flawed assumption that police records are objective reflections of crime, rather than reflections of localized police activity and systemic over-policing. By examining literature on algorithmic fairness and professional ethical codes, such as NSPE 3.1 and 3.3, I identified a critical "technical trap" where engineers attempt to solve social bias through narrow laboratory adjustments while ignoring the political tradeoffs inherent in model design. The research concludes that the lack of transparency regarding data origins constitutes a violation of engineering honesty, advocating instead for a framework that reclassifies predictive tools as instruments for resource allocation, such as housing and social aid, rather than justification for increased surveillance. The technical portion of my thesis produced a multi-stage geospatial pipeline and an interactive dashboard designed to analyze and visualize the structural drivers of Disproportionate Minority Contact (DMC) in the Jefferson Area Criminal Justice System. Moving beyond traditional juvenile-focused studies, my model integrated U.S. Census Bureau data with individual booking records from the Albemarle-Charlottesville Regional Jail (ACRJ) using approximate string matching and geocoding. Through K-Means clustering and Principal Component Analysis (PCA), the project identified that the relationship between race and booking rates is not static but is heavily mediated by socioeconomic context; for instance, a statistically significant interaction showed that the impact of race on arrest rates is drastically amplified in high-poverty census tracts. The resulting dashboard provides local policymakers with a localized visualization of systemic drivers, such as poverty and housing instability, shifting the focus from individual "criminality" to neighborhood-level needs. Having completed both projects, I have learned that technical precision is hollow without a robust understanding of the social context from which data emerges. My technical research enriched my STS critique by providing a concrete alternative to punitive mapping: by mapping socioeconomic stressors instead of just arrest counts, I demonstrated that an engineer can pivot a tool's purpose from surveillance to evidence-based aid. This duality highlights a profound ethical lesson: the "welfare of the public" (NSPE 2.1) is not a default setting but a conscious design choice. My work suggests that when engineers acknowledge the non-neutrality of their data, they transition from being passive facilitators of systemic bias to active architects of institutional trust and social equity.

BS (Bachelor of Science)

Disproportionate Minority Contact ; K-Means Clustering ; Principal Component Analysis ; Data Integrity ; Human-Robot Interaction

School of Engineering and Applied Science Bachelor of Science in Systems and Information Engineering Technical Advisor: Smith Michael, White Preston, Alonzi L. P., Vitale Matthew STS Advisor: Jacques Richard, Laugelli Benjamin Technical Team Members: Jay Sorkin, Mitchell Palmer, Ryan Yates, Kevin Villalobos

English

All rights reserved by the author (no additional license for public reuse)

2026-04-30