When analysing crime patterns, it’s common to rely on timestamps that record precisely when incidents occurred. However, many crime types, such as burglaries, frequently happen within uncertain time windows — for example, a burglary might take place sometime between when the resident leaves at 9 AM and returns home at 5 PM. Assigning the crime solely to 9 AM or 5 PM could significantly distort our understanding of crime patterns. A practical solution to this...
Last Updated: April 20, 2025 - 3 min read
The Home Office Clear, Hold, Build (CHB) initiative is a place-focused intervention that has been happening in England and Wales since 2021. In a recent impact evaluation, analysts used a four-stage approach — combining Propensity Score Matching (PSM), Synthetic Control Method (SCM), Parallel Trends Assessment, and Difference-in-Differences (DID) — to assess whether the intervention effectively reduced crime. Note:This article is not a full review of the CHB initiative, nor is it intended as a primer...
Last Updated: April 20, 2025 - 10 min read
As I rummaged through my Twitter/X account to remove older posts, I stumbled on a thread I thought was worthy of a short article. It concerns how police forces score individuals for targeted action where violence is concerned. The thread was specifically in response to the criticism faced by the Metropolitan Police’s use of what was more widely known as the Gangs Matrix back in 2022. The ‘Gangs Violence Matrix’ (hereafter GVM) was introduced in...
Last Updated: April 20, 2025 - 6 min read
This short post is inspired by the Crime De-Coder blog article ‘A simple rule to identify if rare crimes have spiked’. I thought I would try and do something similar in Microsoft PowerBI as I’ve seen PBI being used for public-facing reports by numerous law enforcement agencies in the UK and elsewhere. Connect to your data. In this example, I directly connect to London homicide data from the Greater London Authority Data Store. You can...
Last Updated: April 20, 2025 - 1 min read
Analysts in policing are frequently asked to prioritise and rank order places and people. Usually, it involves recent data, typically the last 12 months or less, and we expect there will be a disproportionate concentration to focus on. Are these concentrations stable among particular people and places? Are there developmental trends that we can learn from or factor into our selection for targeted intervention? Group Based Trajectory Modelling (GBTM), aka Group Based Trajectory Analysis (GBTA),...
Last Updated: July 10, 2024 - 6 min read
At a time when the safety of women and girls in public spaces continues to be of paramount concern, this post looks to spatial analytical methods and related responses for improving safety. Over the last few years, I’ve completed several pieces of work exploring the safety of public spaces for women and girls. This has included developing my understanding of the risks (victim research and surveys, place-based research), the available data and its limits, testing...
Last Updated: April 05, 2024 - 16 min read
Merging multiple datasets and gaining a view of unique persons in administrative data is a regular obstacle for data professionals. But with Splink that process can be fast, accurate and scaleable. We often need to collate a single-person view in criminal justice and law enforcement. It may be to understand a person’s journey through the system, identify a trajectory or help understand overall risk and harm. When working with legacy systems, it might mean that...
Last Updated: July 10, 2024 - 2 min read
The recent Home Office Evaluation report on Grip and bespoke-funded hot spot policing highlighted differences and challenges encountered by analysts in identifying the best hot spot shapes and sizes for their visible policing strategies. Ideally, we want places small enough and hot enough for police officer presence to generate an effect. This short article offers some thoughts when assessing the suitability of the shapes and sizes of hot spots for visible patrol. This is a...
Last Updated: February 24, 2024 - 9 min read
This week the Home Office published its Evaluation report on Grip and bespoke-funded hot spot policing, which was the largest national effort in observing and evaluating visible patrol as a strategy to reduce crime in hotspots - in this case high violence locations in England and Wales. For similarly large-scale national efforts at hotspots, we probably have to go back some way to the Burglary Reduction Initiative (1998-2002, 247 sites, problem-solving at hotspots) and Street...
Last Updated: February 17, 2024 - 7 min read
If you’re working in law enforcement or with crime data, it’s quite likely that you will have been asked to produce data aggregated to a large spatial unit — like a neighbourhood, census block or ward, or even areas as large as cities and towns — for the purpose of prioritising geographical locations for additional resource But is that an efficient way to target geographically? Perhaps a rationale behind such requests is that large defined...
Last Updated: February 24, 2024 - 2 min read