PredPol, Inc.

PredPol, Inc., later Geolitica, develops predictive policing software that forecasts probable crime locations and times, focusing on property offenses. Its cloud-based platform provides law enforcement agencies with near real-time insights, enabling strategic resource allocation to deter criminal activity and enhance public safety through data-driven prevention. This technology aims to optimize patrol strategies by identifying areas most susceptible to future criminal acts.

The company was founded in January 2012 by mathematician George Mohler and anthropologist Jeff Brantingham. Their collaboration stemmed from academic research, notably applying earthquake prediction models to crime patterns. This insight, initially developed alongside the Los Angeles Police Department, established PredPol’s methodology to introduce scientific predictability into policing strategies, drawing upon the founders' respective expertise in data science and behavioral patterns.

PredPol primarily serves law enforcement agencies, offering tools to optimize patrol deployments based on its predictive analytics. The company's vision sought to empower police departments with actionable intelligence to reduce crime rates. By refining forecasting capabilities, PredPol aimed to transform policing into a more efficient, data-informed practice for community safety and more proactive law enforcement operations.

High-Level Overview

PredPol, Inc., now operating as Geolitica, develops predictive policing software that forecasts property crimes using historical data on crime type, location, and date/time, delivered daily via a secure web interface to police departments.[1][2] The platform serves law enforcement agencies worldwide, aiming to enable proactive patrolling in high-risk micro-areas (as small as 500x500 feet) to prevent crime without relying on personally identifiable or demographic data.[1][2][5] It addresses reactive policing limitations by providing data-driven crime predictions, though its growth has faced scrutiny; as of 2025, the Santa Cruz-based company reports $4 million in revenue and around 4-7 employees, down from broader adoption claims of 60 U.S. departments in 2019.[1][2]

Origin Story

PredPol emerged in 2010 as a research project from the Los Angeles Police Department (LAPD) and UCLA professor Jeff Brantingham, adapting an earthquake aftershock prediction model into a patented algorithm for crime forecasting.[2] Founded formally in 2012 in Santa Cruz, California, the company quickly gained traction, becoming a leading U.S. predictive policing vendor by that year and powering tools used by mid-sized agencies (100-200 officers).[1][2] Early pivotal moments included LAPD integration, where officers received printed maps for GPS-tracked patrols, though independent validation of efficacy remained limited.[2]

Core Differentiators

• Minimalist Data Inputs: Relies solely on three data points—crime type, location, and date/time—avoiding PII or racial/demographic factors, which differentiates it from broader surveillance tools.[1][2]
• Granular Predictions: Generates daily forecasts for tiny zones (500x500 feet), enabling precise patrol targeting via secure web/printouts, with global deployment claims.[1][2][5]
• Earthquake-Inspired Algorithm: Patented model borrowed from seismology for aftershock-like crime "hotspots," positioning it as a novel application of machine learning in public safety.[2]
• SaaS Delivery for Law Enforcement: Focuses on GovTech with emphasis on privacy, though secrecy around client lists and lack of independent accuracy studies set it apart controversially.[1][2]

Role in the Broader Tech Landscape

PredPol rides the wave of AI-driven predictive analytics in public safety and GovTech, capitalizing on big data and machine learning trends post-2010 to shift policing from reactive to preemptive.[1][2] Timing aligned with rising urban crime data availability and post-financial crisis budget pressures on police, favoring efficient tools amid competitors like Palantir and ShotSpotter.[1] Market forces include growing demand for data analytics in law enforcement, but headwinds from bias critiques—termed a "self-fulfilling prophecy" by scholars like Ruha Benjamin due to intensified patrols in predicted zones—have amplified debates on algorithmic fairness in criminal justice.[2] It influences the ecosystem by popularizing predictive policing (most-used U.S. algorithm as of 2020), yet LAPD audits and absent independent proof highlight tensions between tech innovation and accountability.[2]

Quick Take & Future Outlook

Geolitica (PredPol) faces a pivot amid controversies, with stagnant employee counts (4-7) and revenue signaling potential contraction in a post-2020 scrutiny era on biased AI.[1][2] Next steps likely involve transparency enhancements, bias audits, or expansion into non-policing risk analytics to align with ethical AI regulations and trends like federated learning for privacy. Evolving influence may shift from covert adoption to regulated GovTech partnerships, especially as cities demand validated outcomes—tying back to its LAPD origins, success hinges on proving predictions prevent crime without perpetuating inequities.[2]