Data-driven policing requires fundamental changes
In policing, there is no shortage of data. "We are already heading nicely toward the hundred petabytes of data," confirms Oscar Wijsman, International AI & Data Science Lead at the Dutch Police, in response to a poll at the Data Expo 2025. "And that amount just keeps growing." About half the data comes from investigative investigations, the rest from public sources, public order enforcement and business operations. Police work has long revolved around the processing of data, but by searching ever smarter through that vast amount of data, the police can make connections very quickly , take targeted action and work more efficiently.
His colleague Ruud Staijen, national program director Data-driven Working Together, gives as an example the arrest of suspects after the murder of Peter R. de Vries. "The case was solved very quickly because someone saw part of the license plate number of the getaway car. We were able to combine that with other data and link it to route maps. As a result, we had two people 'inside' within half an hour." In such a case, everyone is happy that the perpetrators can be caught quickly, but there is also a very good business case for data-driven policing. "In the past, if we wanted to catch a criminal, we would deploy an observation team. That's pretty expensive. You need as many as 16 men per 24 hours to follow a person or a car. However, this person's phone tells the towers exactly where he is."
Organizational model must change
Wijsman: "All that data affects our way of working. We as police are traditionally organized strongly according to a hierarchical model. For data-driven work, that model is insufficiently suitable and we will have to move to other models. There is a need for more flexibility and working together in a network, for example thinking and working in hub-and-spoke constructions." That switch will have the necessary impact on organizational structure, training practices and how we should work with each other. "Remember, we are an organization of more than 65,000 people. We are the largest government organization after Defense. This is not something you just get done very quickly," says Wijsman. Data-driven work has already created new positions in a short time, which also require a shift in training. "We now have many more colleagues with at least a bachelor's, a master's or even multiple masters. We even have a whole bunch of PhDs walking around with us now."
Still, it is often not clear what is allowed with the data. "I see data as a new weapon for the police officer," says Staijen. "And as with the use of other force, the use must be proportionate. If someone crosses the street diagonally or runs a red light, I don't grab my baton for the arrest either. Every police officer receives training every three months on how that proportionality is defined in laws and regulations. Now I'm going to use data violence on your phone, your Internet behavior or the cameras over the road. How well is that legally regulated? Well, that's still thin on the ground and so we're taught very little about that. We are still on the eve of solving that issue."
Getting started in practice
Because the possibilities of working with data and the legal-ethical frameworks are still developing, the national police are following a strategy with four program lines. One is working with living labs or living labs to address various issues based on practical examples.
A second line focuses on improving the handling of data. Staijen: "Among other things, we did a datamaturity study. I will say: there is room for improvement there." It is far from always clear whether data are correct, who has had access to them and whether the data are still authentic. Especially the metadata, which should be added at the time of submission, is regularly missing. Staijen: "We are doers, which means that we sometimes fail to take care of the data. Then at a certain point you no longer know exactly whose data it is and whether or not it should already have been cleaned up. Actually, that should no longer be part of the normal production process. It becomes another business."
To solve that dilemma, the police chose to establish a datacare organization. This gives police officers advice on what is and is not allowed with data, often in consultation with the prosecutor. Training will be provided to improve legal, technical and ethical knowledge. The datacare organization will also supervise data flows and the growth and use of the data. This will ensure that data governance is in order, and that concern for data quality is no longer an afterthought in the production process.
Legal and ethical professionalization
Further professionalization of the police force with respect to data is a third strategic program line. This line focuses primarily on the necessary professionalism and on strengthening legal and ethical action. Together with the Police Academy, training courses are being developed and a legal network has been set up. This supports the living labs with advice, for example.
A final important line focuses on streamlining all data facilities. A target architecture has been developed to end the fragmentation of data and methods. There are now thirty to forty different analysis tools and various data warehouses in use. Staijen: "Without direction, camps will form that will fight with each other for control of the data. So we are on the eve of a rationalization exercise. That is an expensive word for red lines, cleaning up and standardizing." While doing so, there must remain room for innovation, he emphasizes, so the police organization is looking for a good balance between standardization and flexibility.
Situation Netherlands has advantages
According to Staijen, it has great advantages that the Netherlands formed a national police force more than a decade ago. This allows the transition to data-driven work to happen faster than in countries such as Belgium, France and the United States. "There are not many countries in the world that have one national police force. As a result, you see that data is often fragmented, and there is little willingness to share data between forces within a country."
With the four program lines, the police organization is taking important steps toward a future-proof and efficient approach to security in the Netherlands.