We signed up for the 3rd PNP Hackathon to take on the Predictive Policing Challenge.

We’ve been working on Pantrypoints for most of the year without doing much for ISAIAH, which is the machine learning part of the Superphysics system. So we joined the hackathon to develop the predictive analyis part, which we will call “Criminality Prediction System (CPS)”.

The system uses palm prints to predict a person’s tendency for criminal behavior. Those with such tendencies are called “pre-suspects”. In this way, pre-suspects are monitored by the police before they commit the crime.

This system was supposed to be for the insurance industry in order to ‘predict’ a policyholder’s tendency for injury, health problems, etc.

How is it different from The Minority Report? #

In The Minority Report, the police arrest the suspect before the crime is committed. But with CPS, the police actually allow the crime to happen. They either:

• intervene immediately during the criminal act in order to prevent death or damage, or
• witness the crime directly using their bodycams. This will greatly aid in the trial of the criminal.

We chose Keras over OpenCV because of its deep learning abilities.

Update: October 19 #

We didn’t make it to the Top 10. However, we did learn a lot about:

• The PNP’s mentality
• The justice system with regard to how criminals are assigned
• The many processes in the PNP