Every summer Principle One hosts a team of student interns, typically drawn from a range of different backgrounds and set them up to tackle some of the harder research projects that we don’t have time to tackle ourselves throughout the year. This summer, we welcomed Mathematician Erin Law from Newcastle University, Physics and Astrophysics student Caterina Prior from Lancaster University and Security and Crime Science students, Emily Simpson and Lucy Williams from University College London.
Over the course of ten weeks, we tasked them with exploring the challenge that many of our law enforcement customers face around generating synthetic data for testing, training and exploring new capabilities. The ability to search for digital evidence has becomes more and more critical, whether it’s hidden in a haystack of ANPR data, Communications Data or financial transactions. However, getting representative data to experiment and explore presents data privacy problems and manually creating data can be time consuming and lack realism.
We broke the problem down further; how could we build a synthetic world that reflects the complexity and volume of the digital footprint that each of us creates and how could we inject a scenario based on criminal activity into this world to test out different investigative techniques?
The internship began with a crash course into the world of consulting, with training courses in business analysis and agile working, technical training and a series of increasingly deep dives into the world of digital intelligence, supported by subject matter experts from within our consulting team.
We split our team in two, with Erin and Caterina focused on finding methods to generate two forms of synthetic data; ANPR and Communications Data. This was a daunting task considering the scale and complexity of the datasets in real-life and the complexity of modelling how a real world of road networks, mobile phone masts and Wi-Fi hotspots would work.
Working iteratively, they dived straight into the challenge. Erin reflected “We learnt that we shouldn’t be scared of making mistakes and that getting comfortable with failing fast and questioning our methods as we went would help us get to a better solution.”
Getting the balance right between realism and the limits of the synthetic environment was a challenge throughout. They recognised that different types of vehicles (cars, buses, lorrys) undertook different types of journeys (commuting, taxis, leisure). These journeys also took place at different times of the days, different days of the week, and varied in length and frequency. Assuming each car contained a mobile phone, data was generated as the vehicles travelled along real routes, “pinging” ANPR cameras, cell sites and Wi-Fi routers - outputting data into a series of files based on real life formats that investigators acquire during an investigation. The scale and pace at which we can generate data goes far beyond what has previously been possible and got our subject matter experts across Principle One fairly excited around its potential benefits across our wider customer facing work.
“The initial idea was to build a simple world where phones and cars would generate “real” data as they moved around, just as they do in real life.” said Neil Sumner, Technical Architect and one of the interns’ mentors. “The scale and scope of what Emily, Lucy, Erin and Caterina have achieved in ten short weeks has been staggering and completely exceeded my expectations. They were brilliant to work with, keeping me on my toes with insightful questions and technical challenges. I have no doubt they will all excel at whatever they decide to do next.”
In parallel, Emily and Lucy worked to research the type of data that would aid investigation and the legislation and processes in place to access it. Working against a homicide scenario centered around a County Lines investigation, they were able to create a detailed process map of the investigation, taking into account the decisions that the investigating officer would have to make around where to focus their efforts. For Emily, this provided a real-world perspective to build on her academic experiences at UCL “I loved being able to compare what we learn in our degree against the real-life challenges faced in an investigation and learning more about how an investigation works in practice.”
Once the scenario was developed, it enabled the team to inject the details of the crime scenario, key events and details of the subjects of interest, their vehicles and devices into the synthetic world and create that digital haystack facing investigators. The next step was to start looking at how the data could be visualised and explored interactively.
We called on support from our partners at Elastic to provide a crash course in their Elasticsearch platform which we knew would provide us with the flexibility and scalability we would need, while also being intuitive to get to grips within a short period of time. Lucy took ownership of building a series of interactive dashboards to enable the data to be explored throughout the investigation. This was a new skillset for Lucy and she surprised herself by discovering a passion for data analysis.
Finally, to bring everything together, Emily led the development of an interactive application using Microsoft PowerApps. This was based on an Intelligence Folder that set out the details of the crime committed and tasked an investigating officer with following up the trail left by the suspects, navigating the different data sources and working within the Elastic dashboards to analyse the data, progressing the investigation with each new digital breadcrumb. By seeding the crime scenario into the background noise of the synthetic environment, analysts can search for patterns and potential leads, and if their analysis is successful, progress through the application to the eventual arrest of their suspects, building up digital evidence in their casefile as they go.
For Lucy, being part of the Principle One team has been an important part of the internship “I’ve really loved how closely we’ve been able to work with experts across the Principle One team and it’s kept us motivated throughout the summer to see how excited everyone has been about the project and how much value it could deliver to their customers.”
For Phil Tomlinson, Principle One’s Digital Intelligence Lead and a former Senior Investigating Officer at the Counter Terrorism Command (SO15), the project outputs have surpassed his expectations and could support a step change around progressing data capabilities in policing. This was validated through a presentation to a panel of intelligence and data analysis experts across policing, who have shown a keen interest in the value the project could offer them, whether in training new staff or in experimenting with new AI capability.
“We knew the challenge we’d set our interns was difficult, because as far as we knew, it hadn’t been achieved by anyone else. However, we were also confident that by giving a group of very smart interns the right SME support, the digital tools and the freedom to explore concepts and ideas - they would succeed. This approach, coupled with their natural ability to solve problems and apply complex mathematical theory to a policing problem exceeded our own very ambitious expectations. They will be a very hard act to follow - but we are already thinking how we repeat the experience with another big intern challenge in 2024!”