Big data and predictive analytics have a lot of applications, including sales forecasting, marketing and strategy optimization, machine maintenance, even sports and design. Police and security experts already know that big data and predictive analytics might be extremely helpful in fighting and preventing crimes. We are going to explain how they do it.
In 1883 Alphonse Bertillon created the first-ever biometric maps which helped him to identify 49 criminals listed in 1800 files. It was a pure break through back in the days because the practice of stamping criminals became illegal and the felonies often changed their identity. This is how criminal biometry was born, it was based on the belief that 11 parts of the human body are describing its uniqueness with an accuracy of 1 to 4 191 304. Latter, the Bertillonage was replaced by the fingerprints as more precise identification method but it is still the first case of data utilization in modern policing and crime prevention work.
Fast-forward more than an entire century to the “Minority Report” sci-fi movie starring Tom Cruise, where future crimes were identified before they even happened. Modern life is not so promising but big data and predictive analytics are giving us some promising insights.
As every human activity, committing a crime also has its own patterns. Once identified, such patterns might be discovered in unstructured data and this way helps the police not to be reactive but the opposite. For instance, by using big data you can spot relationships troubled family and another in high-risk potential. If the kids of a high-risk family are recording absences from school, it is time for prevention work.
Video analytics is pretty useful in police work too. A lot of cities are dressed up with surveillance cameras and equipment. In addition to this, some police departments utilize body cams in their patrol work. Big data is already supporting image classification and by combining it with face recognition solution, you turn patrolling officer into effective crime preventing machine. The officer can be noted by the predefined signal that unsuspicious person he just passed by, actually, is a fugitive offender.
Children abuse is another field where big data analytics achievements can be utilized. By using image recognition algorithms, a large amount of graphic content can be filtered and sorted in short periods. This is a valuable outcome for fighting sexual abuse of children. To spot a pedophile creating and trading with such pictures you have to browse gigabytes if not terabytes of data. Big data can do this fast and easy.
An experienced police officer is capable of determining incoming emergency call just by the hour it was made. Anupama Rajam has shared a blog post with data investigation of her own. She took a look at the crime statistics in Philadelphia contributed by Kaggle by OpenDataPhilly. The data set contains crime information over 10 year period and it was broad enough to point out that 2 – 4 pm is the peak of thefts. DUI crimes seem to occur mostly at 1 – 2 am. All other offenses peaks in the hours between 10 pm and 2 am.
Finally yet importantly, is the fraud detection. Modern data analytics has a successful record of accomplishments in this perspective. From banks to e-commerce businesses, fraud detection is minimizing the costs of malicious customer behavior by immediately identifying any suspicious patterns and promptly alarming whoever should be.