Can you imagine the police being able to know when and where a given crime is going to be committed, hours before it happens? Can technology and data analysis end crime in our cities? If you answered no, it may be because you are not familiar with what Big Data is capable of today.
New York City is currently at the forefront with a new trend: applying analytical models capable of predicting where and what time of day a given type of crime will occur. To do so, the New York City Hall (representing the tenth most dangerous city in the world in 2010, according to the World Security Institute) and the Spanish firm Synergic Partners have integrated data from the police themselves (reports), census information (purchasing power, unemployment levels, etc.), news of incidents obtained from The New York Times, stock market values, scheduled events for the city and even weather data.
“We wanted to take it even further, combining all of this data in order to make decisions on whether or not to increase police presence in certain areas and at certain times, in addition to analyzing the evolution of various types of crimes over time and how they might be influenced by factors such as society, the economy and social media,” explained Santiago Gonzalez, Innovation Director at Synergic Partners.
Thanks to this tool, we now know that Brooklyn is the district with the highest crime rate, usually between 3:00 pm and 7:00 pm, with a lower rate in January (perhaps because of the Christmas holidays). According to the model, these crimes are mostly small-scale thefts, with higher frequency on Fridays, when there is also a higher concentration near banks, with four times as many crimes happening in or around them than in other locations.
“One example of information that the police did not previously consider was the number of thefts that happen in the street in front of private residences, with 39.8% of thefts happening on public streets. When we showed them this data, they admitted that they did not have a protocol for systematic patrols,” added Gonzalez. “When we pointed out that crime rates dropped at night, the police responded that they did indeed have more officers dedicated to patrolling those areas than during the day, when they are more likely to be occupied with traffic-related issues.”
With a total of 78 variables and 4.2 Tb of information, the model (actually a set of models in a neural network with decision trees making up a global meta-model) can predict crimes from murder to small-scale thefts, including kidnapping and all sorts of crimes and misdemeanors. “We have achieved a precision average of 72%, with respect to the time a crime will occur as well as identifying false threats,” added the expert. This precision level can be as high as 83% for murders (the type of crime with the greatest amount of available information) and as low as 67% when it comes to predicting a kidnapping. “There are fewer cases of kidnapping and they are normally handled by the FBI, meaning we have less data available,” admitted Santiago.
There is, however, one small catch: the analysis Synergic Partners did on the crimes in New York was carried out a posteriori, given that the city does not provide police information in real time. “Our model is prepared to do it in the moment,” defended the Spanish start-up. That is the great challenge, getting urban areas to provide their updated open data as well as integrating more information sources into the system. “In the second phase, beginning in April with the help of Columbia University, we are going to put sensor systems into police vehicles so that we can see their geo-location at all times.” They also explain that, in the second half of 2017, “we will delve deeper into the data, focusing the areas to be covered from districts to neighborhoods and even residential blocks, with the analysis of over 200 variables including the instants of real murders in order to be able to better detect serial murders.”
We asked Santiago Gonzalez if this type of technology will be used in our country, and his response was bittersweet: “We have a meeting with the Ministry of Interior that has already been cancelled and rescheduled several times. In addition, when we have spoken with them in the past, they have admitted that they are not prepared at this point to put to use all of this big data.” It is therefore easy to understand why Synergic has decided to focus on taking this technology to cities outside of Spain, such as New Orleans or Seattle, at least for the near future.