The Use of Data Analysis to Predict Crime

By 0
The Use of Data Analysis to Predict Crime

A working model for the Ghana Police Service

Information and Communication Technology (ICT) can be applied in many fields including crime fighting and prediction. In Ghana, the Ghana Police Service (GPS) has requested from citizens to assist the service in curbing the upsurge of crime in Ghana. Fighting crime in all forms is commendable but predicting the crime before it happens is the best approach. ICT can be used to effectively predict crime and through timely interventions, help prevent it from happening. As an ICT professional, there is a tried and tested model that the GPS can adopt. This article briefly describes the model and its application.

There is a saying that a proactive police service that prevents crime is better than a retroactive force that responds after the event. This is what forward looking police services or forces across the globe are aiming to achieve. The costs and accompanying benefits in preventing crime outweighs that of responding after the event. The use of ICT to predict crime has come up strongly after the exploits of ICT in the insurance, banking, health, and tax sectors and many countries mostly in the first world are exploiting many ways by which ICT can be used in predicting crime.

New data-driven analyses using ICT can facilitate this kind of proactive work in several ways:

  • they can strengthen decision-making processes on the operational, tactical and strategic levels
  • they can make crime-fighting and prevention more targeted and effective
  • they can ensure improved availability where and when the needs for police services are greatest
  • they can ensure a quicker response through better knowledge of the relevant crime situation

In recent years, this type of analytical work has evolved along with advances in technology and the digital transformation of society. Different police units are adopting so-called predictive analytics to:

  • predict where and when the risk of certain types of crimes is likely to increase
  • predict which people can be linked to an increased risk of committing a crime in the future
  • Predict which areas, groups or individuals have an increased risk as future targets of crime or that the police for some other reason ought to direct their attention to.

There is broad variation in the methods, complexity and practical application of predictive analytics. At the same time, several aspects of predictive policing are supported by the developments in technology:

  • Use of different sources of data. Digitization of datasets renders them accessible to computer analysis.
  • The use of modern ICT to explore patterns and relationships across datasets, as well as to predict how crime is likely to develop in the future.
  • Analyses that provide risk predictions within delimited areas and time slots. This lends the analyses a different operational value compared with long-term trend analysis.

For predictive analytics to be effective, the police must constantly assess which response is best suited to solve a problem and ensure that the necessary measures are carried out and followed up. Analytical work must therefore be closely linked to the tactical and operational work and be firmly anchored in the police’s medium to long-term strategies.

The police can only benefit from predictive policing if the analyses are part of a cyclic working process consisting of at least four elements. These four elements have been tested by the Norwegian Police (Blandhoel, 2015) and can be adopted by the Ghanaian Police service for the same:

  1. Collection and quality control of data: It is important that the datasets are complete, that any potential biases are identified and taken into account, and that the datasets are relevant to the analysis.
  2. Analysis and predictions: Suitable analytical techniques for the topic under investigation are selected, and it is assessed whether several independent analyses should be run in parallel. Uncertainty and validity restrictions are identified and clarified. It is important to understand the underlying methods and reflect on which factors might contribute to crime in a high-risk area.
  3. Operational response and intervention: The analyses and predictions are followed up by a professional police assessment of what constitutes an appropriate response. In some cases this may be visible police presence on the scene; in other cases, usually in collaboration with other players, steps may be taken to alter conditions that encourage crime at the particular location.
  4. Evaluation and measurement of effectiveness: The intervention is followed up by short-term and long-term evaluations of whether the response has had the intended results. It is important that these evaluations are systematic, independent and take uncertainties into account. New data must be generated and analyzed, and if necessary the response must be changed and adapted

In conclusion, predictive analytics has the potential to provide the police with new knowledge, important strategic insight and an enhanced basis for decision making in the police’s preventive and crime-fighting work. However, it is not the analyses themselves that lead to a reduction in crime but the follow ups to the data and analysis through physical interventions.


Author: Samuel Hanson Hagan – (ICT Consultant, Member: Institute of ICT Professionals, Ghana)

For comments, contact author   Mobile: +233507393640.