The London Nighttime Access to Food Retail Options dataset was developed as part of the Data After Dark research project. The study aimed, first, to map access to food retail options during nighttime hours and, second, to integrate this information with the spatio-temporal distribution of night workers in order to identify underserved areas at a granular scale.
This release provides the accessibility (i.e., supply) layer, which integrates open-source public transport schedule data with information on the locations and opening hours of food retail outlets supplied by Green Street (formerly the Local Data Company), which are also available through GeoDS.
The dataset can be further utilised in academic research and policy analysis to assess access to essential services during nighttime hours in London.
Content
The dataset is provided in GeoPackage (GPKG) format and reports travel times (in minutes) to the nearest open food retail outlet across Greater London, measured on a granular hexagonal grid (350-metre edge-to-edge). Accessibility is measured over a continuous 24-hour period, from 6:00 a.m. Friday to 6:00 a.m. Saturday (March 2024). Daytime accessibility (6:00 a.m.–6:00 p.m.), included for comparative purposes, is reported in 3-hour intervals, while nighttime accessibility (6:00 p.m.–6:00 a.m. the following day) is reported at hourly intervals. The data are provided using a bespoke hexagonal grid created by BT, since the distribution of nighttime workers was released at that spatial level.
The first methodological step involved computing travel-time matrices between the centroids of all hexagons for each departure time using r5r, an R package for rapid realistic routing on multimodal transport networks. The routing incorporates GTFS and street network data, drawing on two GTFS feeds: Transport for London services accessed via the Bus Open Data Service and rail services from the National Rail Data Portal. Door-to-door travel times account for walking, waiting, and transfer times. For each departure time, travel times are calculated for every minute within a 30-minute window, with the median value reported (see publication for full parameter details).
In the second step, the travel-time matrices are linked to the Local Data Company Retail Type, Vacancy, and Address dataset. Four retail subcategories were included to provide a broad and realistic representation of food retail availability in London — Supermarkets, Grocers, Greengrocers & Fruitsellers, and Convenience Stores. Using outputs from r5r and the accessibility package, travel times to the nearest hexagon containing at least one open food store are calculated for each departure time, with store availability determined dynamically based on opening hours.
Quality, Representation and Bias
To mitigate boundary effects, food retail outlets located within 2,000 metres of the Greater London boundary were also included. The final dataset comprises 8,337 outlets, representing the most comprehensive analysis of food accessibility in London to date. It should be noted, however, that the analysis relies on a single accessibility indicator—travel time to the nearest open outlet. This measure does not account for the diversity, quality, or density of available food retail options. The objective was to produce an easily interpretable and policy-ready indicator, rather than more complex measures such as gravity-based accessibility models.
The dataset is based on March 2024 data, using GTFS schedules from 1–2 March and the March snapshot of Green Street data. Acknowledging that accessibility varies by day and season, a specific day was selected rather than using averaged conditions. While the dataset reliably captures general accessibility patterns, differences between days, for example those driven by store operating hours, should be acknowledged.
Missing values (NA) are assigned to hexagons that are either inaccessible (i.e., contain no road network) or have travel times exceeding 60 minutes, which is our maximum travel-time threshold.
