Using machine learning to uncover hidden structures and patterns in large datasets of domestic abuse incidents, victims and suspects. Identifying clusters of victims and perpetrators with shared characteristics and behaviours to support more informed and preventative responses.
Why is the research needed?
Domestic abuse (DA) is a national priority, one in five adults may face it in their lifetime. Its effects go far beyond physical injuries, harming mental health, emotional wellbeing, and family life, while also weakening community trust and cohesion at a local level. The annual economic and social cost is around £66 billion, showing the huge strain on public services – it is not just a private family issue, but a major societal problem with wide-reaching damage.
Research shows that families affected by DA are far more likely to struggle with food insecurity. This can happen when someone controls money or resources, blocking access to food, or simply due to poverty – which adds stress and tension at home, often sparking conflict and violence.
What are we doing?
Our project applies machine learning techniques to identify patterns and groupings within large datasets of anonymous domestic abuse (DA) victims and suspects from selected police areas in the East of England. We will use unsupervised methods to categorise incidents, victims, and perpetrators, then integrate this with aggregated geographical and socio-economic data, such as health, care needs, food insecurity, nutrition, and poverty levels. This approach highlights how social and economic factors contribute to DA vulnerability and prevalence, providing a comprehensive system-level perspective to guide local commissioning, enhance support services, and enable more targeted interventions.
How are we working with communities, services and organisations?
This project is a collaboration between two East of England police forces, the University of Essex, and researchers from the University of Hertfordshire.
What will the impact and benefits of this research be?
This research will improve DA support by identifying patterns and clusters in police data on incidents, victims, and perpetrators, enabling better resource allocation and highlighting under-served areas. It will also reveal links between DA and wider health determinants like poverty, food insecurity, health, and care data, helping local services target interventions more effectively for vulnerable communities. Ultimately, these insights will inform evidence-based policies and commissioning to reduce DA risks and enhance victim outcomes.
What do we have planned for knowledge mobilisation and implementation?
We plan to share findings through accessible reports, presentations to police and local services, and workshops with partner organisations to improve service delivery and guide targeted interventions.
Related papers, outputs and resources
This project will produce key outputs including a peer-reviewed paper, a stakeholder report, and presentations to share findings on domestic abuse patterns and socio-economic links.
Who is involved?
- Principal Investigator: Professor Alejandro Quiroz Flores, University of Hertfordshire
- Professor Iosif Mporas, University of Hertfordshire
- Dr Katerina Hadjimatheou, University of Essex
- Emilia Tylenda, University of Hertfordshire
- Rakesh Velayudhan, University of Hertfordshire
Get in contact
Email Professor Alejandro Quiroz Flores at a.quiroz-flores@herts.ac.uk.
