Issues of privacy and policing intersect in concerns about the practices of police forces across the country being increasingly driven by technological innovations. The legality of government surveillance and the threat its spread poses for the future of privacy has been a hotly debated topic in recent history, but another quieter threat to individual liberty lingers. The technological revolution has also brought along sophisticated data-mining systems that can analyze the locations of past crimes and, in doing so, attempt to predict where future crimes are going to happen. Conventional crime analysis has focused on identifying areas in a city that are more prone to crime, commonly called “hot spots.” More recently, the shift has been towards
…show more content…
In “Predictive Policing and Reasonable Suspicion,” Andrew Guthrie Ferguson dissects two different kinds of predictive models used for anticipating when a crime will happen - the “near repeat theory” and “risk terrain modeling.” Near repeat theory asserts that if a crime occurs in one area, it is statistically more likely that the same crime will occur in the same small area again as opposed to somewhere else for a limited time, usually pegged at a few weeks (Ferguson 5). Near repeat patterns are the most simplistic method of predictive analysis - they rely upon simple linear regression models that draw a correlation between one or two factors and one result, instead of taking in a breadth of factors into consideration, and they were commonly used before the introduction of more sophisticated software (Ferguson 6). In the research paper “Predictive Policing: The Role of Crime Forecasting in Law Enforcement Operations,” by Walter L. Perry, Brian McInnis, Carter C. Price, Susan C. Smith, and John S. Hollywood, it is asserted that based on interviews with crime analysts, in the past, “predictive techniques consisted of purely heuristic methods.” Police officers commonly manually identified hot spots on maps that showed pinpoints of locations of crimes, as opposed to relying on maps of hot spots generated by computer programs (Perry et al. 27). This method was relied upon in the CompStat system, and has been proven to be especially applicable and useful when considering cases of residential burglary. Risk terrain modeling, the other kind of predictive model, draws a parallel with the new program that the NYPD is now in the
In “Predictive Policing and Reasonable Suspicion,” Andrew Guthrie Ferguson dissects two different kinds of predictive models used for anticipating when a crime will happen - the “near repeat theory” and “risk terrain modeling.” Near repeat theory asserts that if a crime occurs in one area, it is statistically more likely that the same crime will occur in the same small area again as opposed to somewhere else for a limited time, usually pegged at a few weeks (Ferguson 5). Near repeat patterns are the most simplistic method of predictive analysis - they rely upon simple linear regression models that draw a correlation between one or two factors and one result, instead of taking in a breadth of factors into consideration, and they were commonly used before the introduction of more sophisticated software (Ferguson 6). In the research paper “Predictive Policing: The Role of Crime Forecasting in Law Enforcement Operations,” by Walter L. Perry, Brian McInnis, Carter C. Price, Susan C. Smith, and John S. Hollywood, it is asserted that based on interviews with crime analysts, in the past, “predictive techniques consisted of purely heuristic methods.” Police officers commonly manually identified hot spots on maps that showed pinpoints of locations of crimes, as opposed to relying on maps of hot spots generated by computer programs (Perry et al. 27). This method was relied upon in the CompStat system, and has been proven to be especially applicable and useful when considering cases of residential burglary. Risk terrain modeling, the other kind of predictive model, draws a parallel with the new program that the NYPD is now in the