Using data to better understand crime in cities

Using data to better understand crime in cities

Data61 researchers were able to find the root cause of crimes in cities through data analysis and modelling

Data is not only the key to smart urban planning and development; it can help solve social issues in cities, too. A team of researchers at Data61, in collaboration with Harvard University, are using data for just that – to get to the root cause of crime in cities.

For a long time, our understanding of crimes in cities has been this: The bigger the population, the more crime. But this is too simplistic of a view, says Manuel Cebrian, computational social scientist at Data61.

His research looks beyond that, as it finds a statistical link between the rate of completed or successful crimes and the size of the police force, and the population size.

Cebrian and his team analysed and modelled FBI datasets, crime counts, population statistics, and police force data for 1,994 cities and towns in the United States.

What he found was not that there are more criminals in larger cities compared to smaller ones, but rather the higher rate of completed or successful crimes in bigger cities was a direct result of deficits in the police force.

The research found that the bigger the city, the smaller the police force, even though it had almost the same amount of crime proportionally to smaller cities.

“For a while, scientists studying cities thought that cities somehow would attract more criminals, somehow criminals would be better off in a city or the social economic features of cities would make criminals thrive.

“But we saw that wasn’t the case. What we did was take all of the data that was available for both crime that was only initiated and crime that was completed. The number of crimes that are initiated is proportional to the size of the city. This means there are not more criminals in larger cities than smaller cities - there are more people, and therefore the percentage of criminals is constant.”

However, the real issue is bigger cities do see more crimes unresolved or not being prevented in time, which might be the reason for the ‘big city means more crime’ stigma, Cebrian said.

The data showed that the average time from initiation of a crime, to police detecting it, to then getting to the crime scene or criminal to resolve it took significantly longer in bigger cities than it did smaller ones. The size of the police force in cities decreased as the population size increased, meaning the root cause of the problem was the police staff count was disproportionate to the population size of the city.

“You could actually just do counts on both sides. But the counts on both sides do not really reflect the underlying model because there is a dynamic between criminals’ planning and executing of a crime and the police detecting and deciding to go after them. So we need to have a model that accommodates these time scales well.

“Every single city is different, every single local police force is different, but there are some commonalities that can be captured in the time scales and the way that the police is addressing the crime.”

In the long term, Cebrian hopes to develop his research and model to be able to assist city councils and governments. He said city councils and governments for a long time have been doing simple counts and metrics around crime.

But they need to go beyond this and do more sophisticated data analysis and modelling to ensure they are addressing the root cause of problems and not basing their understanding on long held views or assumptions, he said.

“Simple counts on police force and crime is missing this dynamic between the initiation and the response of the police force. The model that we are using is to understand the interaction between criminals and the police.”

Join the CIO Australia group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.

Join the CIO newsletter!

Error: Please check your email address.

Tags dataData61statistical modellingsmart citiesdata analysis

More about FBIHarvard University

Show Comments