Predictive Analytics Saves Lives – it really does!

Picture this: I read the overnight incident log and it contains a horrible murder of a vulnerable person in the community. My first thought is for his immediate family and friends. My second is to see if there was any chance of preventing it. If so, what can we learn? Most serious case reviews show that a lack of cross-agency information sharing and timely decision making directly contributes to these sad outcomes. Looking at this case as part of the serious crime review process revealed similar issues.

A failure to identify patterns and risk escalation

What were the facts? The offender was a schizophrenic who, over time, had repeatedly threatened the victim. The last few weeks showed an escalation of offending where the victim was subject to numerous ‘threats to kill’. Threats to kill are not unusual – especially in inner city areas. On the day of the murder, a decision was made to release the offender from a secure hospital without the complete information being provided. What was pertinent to me was the recency, frequency and gravity of the offenses carried out. Also notable was a multi-agency model where incidents can be treated in a transactional and isolated way – not connecting the dots. This tactical approach is unfortunately all too common; to improve operations, a more strategic approach that accounts for the scale and scope of multiple agencies should be considered.

Context needs to be understood

Public sector resources are tight. Years of reduced budgets, increased demand, and elevated risk levels have contributed to a stretched and challenged operating environment – both organizationally and personally. If you get things wrong – people die. You can even go to jail.

Think about this: on any given night, the command and control screen could be filled with 300+ incidents that require police action. For example, Avon and Somerset Police has thousands of people connected to offences. It’s impossible to make sense of the blizzard of information coming from multiple data sources using traditional methods…

The small number of officers on duty will often be dealing with serious offenses—rapes, assaults, missing children, burglaries in progress and kidnapping, as well as serious and life changing road collisions. They are busy…really busy. They work hard and put their own lives at risk every time they don the uniform. They live to protect you from danger when you most need them.

But they are human, and the system has forced a model that is reactive and not proactive. If you only respond when things go wrong, you will always be dealing with things that go wrong. It’s a vicious cycle that does not improve societal outcomes. Of course, preventative work does happen, but there is not enough. Things can easily be missed. The question is, what can we do to ensure things change?

The moment my jaw dropped

Let’s go back to the murder. After reading the incident log, I went straight to my ‘in testing’ Qlik Offender Management App. The app was developed to identify offender risk and escalation using a predictive and data-driven approach. Offender scoring was richly put into context through many data sets visualized throughout the app.

And guess what? The offender on this case was at the top of the risk table! Not only was he at the top, but he also was indicated as escalating in risk. The recency, frequency and gravity of the situation was starkly brought into context with the visualized data. I was stunned and shocked.

So, what does this mean for the public?

The public deserves—and demands—action. I immediately rolled out this app throughout the organization, and to this day it is being used to prevent and reduce harm in the communities of Avon and Somerset. The app supports a data-driven approach to understanding risk, the escalation of risk, and most critically the prioritisation of tackling the most dangerous people at the earliest point. Professional decision making remains paramount – the app does a lot of ‘leg work’ and supports better decision making – but the human element remains a critical piece of the equation. Crucially, it also provides a safety net as it churns away 24/7.

Countless lives have been saved as a result, and much harm has been prevented. Chief Constable Andy Marsh recently talked to the BBC and said the following:

“The apps should help prevent future cases similar to the killing of Bijan Ebrahimi, a disabled Iranian refugee who was beaten to death on a Bristol estate in 2013.”

Check out the full BBC article here.

Using Qlik at scale, democratised to 3,500 officers and staff, Avon and Somerset have transformed traditional ways of thinking – driving efficiency and saving lives. I am proud to continue the charge and support public sector and healthcare colleagues around the world in their journeys. That’s why I jump out of bed everyday.

Our own @datanostic_uk takes a deep dive into taking a data-driven approach to understanding and mitigating risk. See how the Qlik Offender Management App is empowering police to save lives:


In this article:

Keep up with the latest insights to drive the most value from your data.

Get ready to transform your entire business with data.

Follow Qlik