Homelessness Prediction Model

The homelessness prediction tool was developed to foresee how a rise in homelessness would hit areas across the West Midlands after the ban on bailiff-enforced evictions in England ended in May 2021.

Ongoing pandemic and enforced restrictions significantly changed normality in 2020. In terms of forecasting it implies that the main assumption of many statistical techniques – that existing patterns will continue in future – has been severely violated and the use of historical data has been limited. The prediction tool therefore relies heavily on what was known about likelihood of people to experience housing problems before the pandemic onset. In addition, it intertwines this knowledge with recent data on how many people were stricken by the crisis and its ripple effects. The analysis is done at Parliamentary constituency geographical level and is supplemented with DWP statistics on people on Universal credit at postcode districts level.


Research Theme 5

Regional Society and Communities


The homelessness prediction model is an interactive dashboard targeting early preventions of homelessness and reducing numbers of households that are evicted due to rent arrears. The model uses high-level publicly available data, such as the number of Universal Credit claimants, to help identify geographical hot-spots and areas at higher risk of eviction from the private rented sector.


Research Team

Dr Maryna Ramchan (lead) 

Paul Muir, WMCA Homelessness Taskforce Associate Support 


Project lead contact details: 

Dr Maryna Ramcharan

Project support contact details: 

Matthew Patterson, Centre Manager, City-REDI / WMREDI 

WMREDI is funded by Research England and the WMREDI partnership