Fighting Crime with Data: Can ‘Predictive Analytics’ Prevent Crime Before It Happens?

Introduction

Police organizations globally are facing a tough task with keeping crime rates down with less personnel and resources. Many police organizations around the globe are advancing with new crime-fighting techniques with an innovative 21st century approach. The growing use of predictive analytics technology is preventing criminal activity by more accurately by targeting investigations, deploying personnel and allocating limited resources. Predictive analytics software play a key role in helping law enforcement successfully forecast criminal activities and deploy resources effectively in line with community expectations, while decreasing crime and improving public safety. By using predictive analytics data mining, text mining, data collection, and statistical analysis police organizations worldwide are able to better understand and predict future criminal behaviour by analyzing large amounts of data including; thousands of incident reports, tip offs on crime, calls for service and criminal databases, as well as data gathered through citizen feedback. Predictive policing using data from various sources will definitely be a positive step and provide improved direction to help position officers in high risk zones. Predictive policing is taking data from different sources, analyzing them and then using the results to anticipate, prevent and respond more effectively to future crime. It entails becoming less reactive. The predictive vision moves police organizations from focusing on what happened to focusing on what will happen and how to effectively deploy resources in front of the crime, thereby changing its outcomes. The data itself is not especially useful until it is structured, interpreted and analyzed. At that point, because of the volume of data available, it becomes valuable in identifying patterns and predicting trends. The use of this data to forecast future behaviour is known as “predictive analytics.”

Predicting Crime Before It Happens

Predictive policing seeks to forecast where and when crime will happen and intelligence focuses on who will commit crime or become a victim. Data analysis in policing has been around since 1995, when the practice of analyzing crime data for geographic patterns to identify crime “hotspots” was first enabled by the introduction of CompStat. A management model, CompStat tracks crime locations and case statuses while enabling a primary level of analysis. CompStat provided management intelligence to law enforcement which included basic forms of the advanced crime maps and predictive models employed across police organizations today. CompStat, however, was not updated with real-time information, it relied heavily on human interpretation of data. Many improvements in technology have since made data science one of the most important tools for police organizations. In the 2003 science fiction film Minority Report has already been realized. The film portrayed police using superhuman “precogs” to predict crime, allowing them to arrest the individual before the act could be committed. Imagine a world in which we could predict crime and police officers would be sent to the incident spot as the crime is happening, or even before it happened. This could be achieved by predicting situations based on data trends in crimes. By analyzing data for serious crimes such as woudings/shootings, murders and robberies, patterns and trends in the behaviour of these crimes can be noted down which would in turn help in reducing these crimes. It makes tasks easier for police by adding information into the system and for information to be provided in real-time to front line officers to help them focus more effectively on prevention of crime. Advanced analytic capabilities can be integrated into CCTV systems to improve response times to public safety incidents. Analytics has the potential to enable police officers to achieve a truly preventative approach, as it can help to get new understandings from data and identify and recognize suspicious behaviour and activities. 

Uncovering Patterns to Predict Crime

Imagine a world in which we could predict crime; a world where police officers would be sent to an incident before it even happened. Police officers are first responders to emergencies large and small, including natural and man-made disasters. Among their varied responsibilities, the reduction of crime is one that police leaders continuously strive to find improved methods by which to be successful. Reducing crime reduces social harm, improves community resilience, and strengthens community security. Improvements in technology and data collection have opened new avenues for police executives to pursue in making improved crime control strategies. The use of mathematical algorithms to analyze data and predict where a crime is likely to occur is becoming more common in law enforcement. Police organizations are battling the exponential growth of data such as community feedback, criminal arrest records, and crime patterns, and the task at hand is to turn large and disparate volumes of data into actionable information. Predictive analytics data-mining software uncovers patterns in data using predictive techniques that play a critical role in decision making. By using predictive analytics, those organizations are able to quickly analyze massive amounts of incident data along with current and developing conditions. Text mining is another component of predictive analytics that leverages critical insights locked in unstructured data, such as incident reports and witness. Text-mining software to can uncover hidden patterns and relationships in text by creating an automated index out of the unstructured contents of a PC’s hard drive, enabling police investigators to perform keyword searches for evidence. Analyzing data allows police organizations to confirm what they already know and discover new information. In many cases it has become a critical tool in the ability to anticipate crime. Predictive policing never tells you exactly what’s going to happen, but it tells you there is a high likelihood of an incident based on prior events. Predictive analysis is in use today in many police organizations as part of a strategy of predictive policing. The accuracy of the analysis is dependent on the quality and quantity of historical data available. Police leaders should embrace the reduction of community fear of crime as part of their organizations missions and incorporate evidence-based approaches to reducing fear of crime.

The emphasis on doing more with less, a mantra for police organizations around the world, appears to be a policy that is required. Other parts of the world are following suit:

• Asia. Advanced analytic capabilities have been integrated into CCTV systems to improve response times to public safety incidents.
• Africa. Police are beginning to build crime prevention into their strategic approaches, through better crowd management and police dedicated to reducing road traffic accidents.
• South America. Police are tackling border protection issues by using technology to better monitor federal highways and manage traffic. 

Police organizations need to be given better access to data that will drive actionable intelligence and free up time to concentrate on investigating crime on the front line rather than dealing with time-consuming administrative tasks behind the scenes. Analytics has the potential to enable police to achieve a truly preventative approach getting new understandings from data, identifying and recognizing suspicious behaviour and activities, and enabling officers to get a head start on the criminals.

Examples of Analytics Being Leveraged For Crime Prevention

The Indian Police force has started taking an increasing interest in crime analytics using big data, which involves storing and analyzing huge volume and variety of data in real time, to predict and inference patterns and trends especially relating to human interactions and behavior. To know which areas are most prone to crimes, the police force also uses predictive analytics to develop models using machine learning to know which areas are most prone to crime. It also helps them to keep a track on which criminals or individuals to keep a track on. Delhi police have partnered with ISRO to develop an analytical system called Crime Mapping, Analytics and Predictive System (CMAPS ), which helps the Delhi police to ensure internal security, controlling crime, and maintaining law and order through analysis of data and patterns. Jharkhand police force is trying to implement an analytical system with the help of IIM Ranchi, to help evaluate criminal records, date and time of crime occurrences, and location to predict crime-prone zones. The system is built on machine learning algorithms and behavioral science and intends to solve crimes all over the country. 

The Bangalore police with the help of IBM is now training officers on the better usage of data analytics software. The Indian police force has started taking an increasing interest in crime analytics using big data, which involves storing and analyzing huge volume and variety of data in real time. Reading Police Department leaders of the US had recently unveiled a crime-reduction strategy called CompStat that embraces analytics with the vigor that many sports teams today use to get a step ahead of the competition. CompStat has been credited with significant reductions in serious crime in New York City and has been adopted by numerous police departments today. The South American police is tackling border protection issues with the help of analytics to better monitor federal highways and manage traffic. Africa is beginning to build crime prevention into their strategic approaches, through better crowd management and police dedicated to reducing road traffic accidents.

Big Data Analytics and Crime Analysis

The world is an increasingly connected place that continues to produce more and more data. Phone numbers, emails and financials can be studied for suspicious links. Government organizations collect health, educational and criminal records. Police can monitor Facebook, YouTube and Twitter feeds. This is the big data world of law enforcement offering much more incriminating pieces of data to use, surveil and investigate. Behind this data is technology: algorithms, network analysis, data mining, machine learning and a host of computer technologies being refined and improved every day. Police can target the crime networks while analysts can link suspicious behaviour for further investigation. The decisional work of identifying criminals, crime networks and patterns now starts with powerful computers crunching large data sets instantaneously. A key feature of advanced analytics is the use of algorithms, which increasingly incorporate Artificial Intelligence (AI) methods underpinned by machine learning. Predictive policing introduces a scientific element to law enforcement decisions. The benefits of data analytics in policing are huge with the benefits of analyzing data from body cameras, sensor networks and smart devices. The data police organizations collect is usually unstructured. Datasets tend to be stored in separate data repositories and collated in ways that make connecting criminal networks difficult. In fact, manual data collection remains common across many police departments locally. To deepen this problem further, lack of standardization is endemic across many government agencies locally as well. Police organizations need access to the expertise to make sense of the volumes of information compiled. This includes structured and unstructured data, such as images and videos. However, getting the right personnel to link these data sets and unveil the “unknown unknowns”, remains to be difficult. Any technology solution that provides police organizations access to their data in an easy to read format, available in near real time and on a range of devices will help with any investigation. The Los Angeles Police Department (LAPD) has set up a Real-Time Analysis Critical Response (RACR) System in collaboration with an IT firm, Palantir . The system incorporates a wide array of data like age, description, address, tattoos, gang affiliations, and vehicle ownership among many others. With just the first name and a physical description, the big data software can narrow down the suspects to a mere few. Automatic License Plate readers that use Natural Language Processing techniques to get the location of vehicles from live surveillance cameras can track down a vehicle in less than a minute. Using big data, LAPD generates crime heat maps daily that are a prediction of possible crimes by utilizing the tons of historical data and real-time indicators that the LAPD stores.

Looking Forward

Technology is constantly changing and improving and predictive analytics is just one way law enforcement can do better with less resources. Police, along with data, can solve and even prevent these crimes by the use of accurate and timely data of crimes. Predictive Analytics can help police with their job and even come up with better results. Though criminals will try to be one step ahead of the law, police organizations by deploying predictive analytics are able to maximize the effectiveness of their personnel and other resources. The key to success in predictive policing is getting as much data as possible to determine patterns. Predictive Analytics promises to improve police efficiency by generating leads about where crimes may occur prior. The results are not precise, but predictive analytics can guide officers toward places most likely to need their help. Big data and predictive analytics the cornerstones of analyzing multiple data types and data sources to predict and even prevent crimes from happening. 

References

Bousquet Chris. April 26, 2018. Mining Social Media Data for Policing, the Ethical Way. Retrieved on March 27 2019 from: https://datasmart.ash.harvard.edu/news/article/mining-social-media-data-policing-ethical-way

Edwards Tom. November 26 2015. How predictive analytics are helping to fight crime. Retrieved on March 27 2019 from: https://www.bluefinsolutions.com/insights/tom-edwards/november-2015/how-predictive-analytics-is-helping-to-fight-crime

Heinze Justin. Fighting Crime with Data: How Law Enforcement is Leveraging Big Data Analytics to Keep Us Safe. Retrieved on March 27 2019 from: https://www.betterbuys.com/bi/fighting-crime-with-data/

Dragosavac Goran. (2013) Big Data Analytics. Retrieved on March 27 2019 from: http://www.bigdatanalysis.com/crime-prediction/

Envisage Technologies. April 1, 2014. Law enforcement agencies use predictive policing to fight crime. Retrieved on March 27 2019 from: https://www.envisagenow.com/law-enforcement-agencies-use-predictive-policing-to-fight-crime/

Das Sejuti. March 21 2017.How predictive analytics helps Indian police fight crime. Retrieved on March 27 2019 from: https://www.computerworld.in/feature/how-predictive-analytics-helps-indian-police-fight-crime 

Hvistendahl Mara. September 28 2018. Can ‘predictive policing’ prevent crime before it happens? Retrieved on March 27 2019 from: https://www.sciencemag.org/news/2016/09/can-predictive-policing-prevent-crime-it-happens

Zimmer Adrienne. March 01 2018. Fight Crime with Big Data and Predictive Analytics: Predictive analytic technologies can help identify where crime might occur and how to respond. Retrieved on March 27 2019 from: https://www.officer.com/command-hq/technology/computers-software/article/20988220/data-analytics

Richey Erin. June 3 2014. How Data Analysis Helps Police Departments Fight Crime. Retrieved on March 27 2019 from: https://www.forbes.com/sites/emc/2014/06/03/data-analysis-helps-police-departments-fight-crime/#2224032d3952