Enlighten Research Data

In this section

Home locations of mobile users for 55 major UK cities (2017-2024)

Li, L. and Sinclair, M. (2026) Home locations of mobile users for 55 major UK cities (2017-2024). [Data Collection]

Collection description

Overview:
This dataset provides annual estimates of home locations for anonymised mobile users in 55 major UK cities from 2017 to 2024. For each unique mobile device (represented by an anonymised user identifier), the most likely home area is inferred on a yearly basis and recorded at neighbourhood statistical geographies: Lower Layer Super Output Areas (LSOAs) in England and Datazones in Scotland. Each estimated home area is enriched with additional spatial referencing (e.g., Local Authority and MSOA/Intermediate Zone codes and names) and linked to area-level socio-demographic context, including deprivation indicators derived from IMD 2019 (England) and SIMD 2020 (Scotland). Temporal coverage includes full calendar years for 2017-2023 and January-June for 2024. The data product is designed to support downstream spatial and social analysis of urban populations while preserving user anonymity.

Estimating a mobile user's home location area involved the following steps:
1. Accuracy filtering: only retain mobile location points with a reported spatial accuracy of ≤100 metres
2. Residential land-use filtering: only retain points located within residential or mixed residential building footprints (Geomni's UKBuildings, see below).
3. Evening activity identification: select points occurring during evening hours (20:00–06:00).
4. Spatial assignment: spatially join each qualifying point to its corresponding neighbourhood geography (LSOA in England/Datazone in Scotland).
5. Active evening aggregation: for each user and area, calculate the total number of active evenings.
6. Home area selection: select the area code with the maximum number of active evenings as the home location for that year;retain only user-year records with at least 2 active evenings in the home area as a single evening is considered too uncertain; if multiple areas are tied for the maximum, exclude the user-year record to avoid ambiguous assignment.
7. Attribute enrichment: join deprivation indicators and higher-level geographies using the estimated home datazone/lsoa code.

Huq dataset:
Huq is a mobile phone app dataset. The app collects real-time, anonymised location data from users' smartphones, based on the use of a range of smart phone applications. This dataset offers researchers insights into human mobility patterns and behaviour. Researchers can leverage this data to study consumer trends, urban planning, and the impact of events on people's movements, amongst other applications. Huq data offers the potential to understand changing societal dynamics and make data-driven decisions across various fields, from retail and transportation to public health and urban development. The data has geographic coverage across the UK

Access and restrictions:
Home locations of mobile users for 55 major UK cities (2017-2024), as well as other derived datasets based on Huq data, were available for non-commercial academic research use only.

More information:
Note:
1. Residential land use data source: Geomni’s UKBuildings dataset (October 2024), a national multi-polygon spatial dataset representing the use and extent of buildings across the UK. Residential and mixed residential buildings are used to constrain candidate home locations.
2. Home locations are estimated independently within each city, so a user may appear multiple times in the same year across different city files (one record per city). Such cases can be identified by the same user identifier (uid).
3. The user identifier (uid) can also be used to join this dataset with the corresponding uid field in the Transformed dataset (title TBC), enabling analysis of mobile user activities associated with home locations.

Funding: