Predictive Policing: Harnessing Technology for Improved Public Safety

Predicting crime is a bit like predicting the weather — you never know what’s going to happen, but some patterns and indicators can give you a reasonable understanding.

When it comes to this, there is no case too big or too small for technology to get involved, because maybe, the time has come to rid ourselves of deerstalker hats and rather equip our precincts with current-day tools and ideas.

One such idea which has been proposed and also implemented is predictive policing. Predictive policing is the use of data analytics and algorithms to identify patterns of criminal activity and predict where crimes are likely to occur.

By identifying areas where crimes are more likely to occur, police departments can deploy patrols and other resources more effectively, potentially deterring would-be offenders. This can be especially useful in cases where crimes are committed by repeat offenders or in specific locations.

So what are the tools available as of date to assist in predictive policing?

One could always sit for hours and manually go through stacks of files and databases to fit each piece perfectly into place until a pattern is identified. But this process requires heavy labor and much time. Machine learning is one of the best available tools to detect crime patterns. Using historical crime data ML algorithms have the potential to predict future crime.

Consider the instance of Series Finder, an algorithm developed by researchers from MIT and the Cambridge Police Department. The algorithm searches a database extensively for patterns till it identifies a modus operandi(habitual way of operating) of an offender or group. It assumes that each modus operandi is different. For example, some offenders operate during weekdays and target apartments while other thefts are targeted at weekends in single-family homes.

The algorithm was tested on a dataset of 4855 households, for 10 years. As Series Finder grows the pattern from the database, the modus operandi for the pattern becomes better defined. The results of the tests proved to be fruitful as they managed to narrow down the suspect description and recover additional crimes within patterns unknown to the police department.

While such algorithms are a breakthrough in the field of ML and law enforcement, there are concerns regarding the potential for bias in the algorithms and the data used to train them. . If the data used to build the models is biased, the predictions made by the models may be biased as well. This could lead to over-policing in certain neighborhoods and under-policing in others.

Policing is a sensitive topic of discussion, especially in recent years. The use of technology must be taken with a pinch of salt and leave no room for prejudice. Law enforcement agencies have to remain open to the idea that while some problems can be resolved with assistance from computers some need to be handled keeping in mind the impact they have on privacy, civil rights, justice and other ethical factors.