UK Biobank Urban Morphometric Platform (UKBUMP) – a nationwide resource for evidence-based healthy city planning and public health interventions

Figures & data

Figure 1. Gene, environment and the corresponding urban health niche for cardiovascular disease.

Figure 2. Spatial locations of 22 UKB assessment centres with number of participants.

Table 1. Description of BE morphological metrics of the UK Biobank Urban Morphometric Platform (UKBUMP).

Urban morphometrics	Data source	Description	Spatial analyses conducted	Spatial scale
Land-use morphometrics
Density of health-promoting/-inhibiting destinations	UK Ordnance Survey AddressBase Premium^1,2, National Public Transport Access Nodes^3 (NaPTAN) dataset	Densities of more than 200 categories of potential health-promoting/-inhibiting destinations were measured within predefined street catchment area. These included agriculture (4)*, community services (12), education (17), hotels (4), industry (8), leisure facilities (24), medical facilities (10), animal centre (6), offices (2), retail (10), transport services (14), bus stops (1), utilities (10), emergency and rescue services (7), information centre (4), allotment (1), amenity land (5), parks (5), unused land (2), water body (4), military sites (6), parking (1), garage (2), house in multiple occupancy (4), residential institutions (4), monuments (6), underground features (2), places of worship (9)	Component layers of OS AddressBase Premium data were joined together through the unique field –UPRN. The exact location (geo-referenced grid coordinates), land-use classifications and full address of each class of health-promoting/-inhibiting destinations were extracted. Density was calculated as the number of units divided by the participant’s predefined catchment area	500, 1000, 1500 and 2000 m street catchment of an UKB participants’ dwelling and census-defined LSOA Developed for all UKB assessment centres
Health-specific destination accessibility	UK Ordnance Survey AddressBase Premium	Network proximity to 36 health-specific destinations including public/village hall/ community facility, job centre, college, children’s nursery/creche, preparatory /first/primary/infant/junior/middle school, secondary/high school, university, factory/manufacturing, mineral/ore working/quarry/mine, workshop/light industrial, warehouse/store/storage depot, library, bingo hall/cinema/ conference/exhibition centre/theatre/concert hall, dentist, GP practice surgery/clinic, hospital/hospice, central government service, local government service, bank/financial service, retail service agent, post office, public house/bar/night club, restaurant/cafeteria, fast food outlet/takeaway (hot/cold), bus stops, car/coach/commercial vehicle/taxi parking/park and ride site, station/interchange/ terminal/halt, electricity sub-station, landfill, power station/energy production, water/waste water/sewage treatment, recycling, police/ transport police station, fire station, ambulance station, places of worship	Street network distance (in metres) between the UKB participants’ dwelling and the nearest health-specific destination was calculated using closest facility analysis in Network Analyst, ArcGIS 10.2	UKB participants’ dwelling unit Developed for all UKB assessment centres
Accessibility to food environments	UKMap data^4 for London	Densities and street network distance to nearest were measured for 19 different typologies of food-related destinations including supermarket, general convenience store/grocer/mini market, butcher, greengrocer, fishmonger, baker, delicatessen/specialist, confectionary, off-licence, restaurant, take-away, coffee shop/internet café, sandwich bar, public house/wine bar, hotels/guest houses/bed & breakfast, department store, general store (WHSmith, Poundstretcher, etc.), newsagent, markets/mixed bazaar	Seven layers of UKMap were linked together through a unique Geographic entity type sequential number (GTN) to extract location (geo-referenced grid coordinates), addresses and precise land-use classes. Food-related land-use destinations extracted. Density was expressed as the number of units divided by the participant’s pre-defined catchment area, while network distance (in metres) between the UKB participants’ dwelling and the nearest was in Network Analyst, ArcGIS 10.2	500, 1000 and 1500 m street catchment of an UKB participants’ dwelling and census-defined LSOA for participants of UKB assessment centres of BARTS, Croydon and Hounslow
Urban network morphometrics
Street-level accessibility	UK Ordnance Survey ITN^5	Urban network analysis model called sDNA was employed to measure the street-level physical accessibility. A suite 20 accessibility indices including network centrality (closeness and betweenness centrality), network detour (sum of crow flight, mean diversion ratio and diversion ratio), network shape and efficiency (convex hull area, perimeter and bearing, and network shape index), link characteristics (length, angular curvature and connectivity) and radius-based indices (number of links, total network length, total angular distance, total and mean geodesic length and number of junctions within a defined catchment radius) were measured	The OS ITN layer was extracted and subjected to automated cleaning including the initial processes of removal of traffic islands as well as repairing of split links. sDNA modelling enabled measurement of physical accessibility indices at pre-defined catchments for each street link in the urban network of the study area. Spatial linkages in ArcGIS between the sDNA network metrics and the dwelling location of the UKB participant enabled measurement of physical accessibility indices of the street network link closest to, within a 25 m buffer and within a 50 m buffer of the UKB respondent’s dwelling	Nineteen different catchment radii from 400, 800, 1200, 1500, 2000, 3000, 5000, 7500, 10,000, 12,500, 15,000, 17,500, 20,000, 25,000, 30,000, 35,000, 40,000, 45,000, 50,000 m to account the influence of physical accessibility at micro-(neighbourhood), meso-(city) and macro-(regional) levels Developed for all UKB assessment centres
Dwelling-level morphometrics
Residential density	UK Ordnance Survey AddressBase Premium	Densities of nine categories of dwellings (including caravan, detached, semi-detached, terraced, self-contained flats, sheltered accommodation, etc.) were measured within pre-defined street catchment area	The exact location (geo-referenced grid coordinates); land-use classifications and full address of units under each dwelling category were extracted from OS AddressBase Premium data. Density was calculated as the number of units divided by the participant’s predefined catchment area	500, 1000, 1500 and 2000 m street catchment of an UKB participants’ dwelling and census-defined LSOA Developed for all UKB assessment centres
Building age and type	Cities Revealed building class dataset^6	Six building age categories comprised of late Victorian/Edwardian (1870–1914), World War 1–2 (1914–1946), post-war regeneration (1946–1964), sixties/seventies (1964–1978) and modern (1978–1988)Twelve building type categories comprised of very tall flats, tall flats (6–16 storeys), lower 3–4 stories and smaller flats, tall terraces (3–4 storeys), lower terraces (2 storeys with T-rear extension), lower terraces (small), linked and step-linked houses, planned balanced mixed estates, standard size semis, semis in multiples of 4/6/8, large property – semis, smaller detached houses, large detached houses	The building class data were extracted and the corresponding footprints were linked with the geocoded UKB participants’ residences. Building age and type was extracted through a spatial query	UKB participants’ dwelling unit Developed for all UKB assessment centres
Physical environment
Greenness	0.5-m resolution colour infrared image^7 (CIR) licensed by Blue Sky	Degree of greenness was calculated as the NDVI (for 0.5 m × 0.5 m) raster cells	A collection CIR data chunks were merged together, area of interest extracted. NDVI index was calculated in Raster Calculator – Spatial Analyst, ArcGIS 10.2. Greenness was calculated in terms of mean, minimum, maximum and standard deviation in the NDVI values within a predefined catchment area	500 m Euclidean buffer of an UKB participants’ dwelling Developed for all UKB assessment centres
Terrain	5-m resolution DTM^7 licensed by Blue Sky	Terrain and degree of slope variability	A series of Blue Sky DTM data chunks were mosaicked together and the area of interest extracted. Slope analysis was conducted in Spatial Analyst, ArcGIS 10.2. Terrain and degree of slope variability (in degrees) were expressed in terms of mean, minimum, maximum and standard deviation values within a predefined catchment area	500 and 1000 m Euclidean buffer of an UKB participants’ dwelling Developed for all UKB assessment centres
Area-level deprivation
IMD	English/Scottish/WIMD	Degree of deprivation expressed as an overall deprivation score as well as with respect to deprivation sub-domain scores for income, employment, health and disability, education, access to services, community safety/crime, physical environment and housing) for 2008 and 2010/2011	The IMD data were spatially linked to the LSOA boundaries obtained from the Office of National Statistics. They were subsequently linked to geocoded UKB participants’ dwelling to extract the LSOA-level IMD scores	LSOA of an UKB participants’ dwelling Developed for all UKB assessment centres
Individual and neighbourhood wealth
Property price	UK land registry data for Wales	Property price in pound sterling and average property price within the catchment	Property price data were geocoded and linked to geocoded UKB participants’ dwelling and individual and neighbourhood wealth calculated	UKB participants’ dwelling unit and postcode of residence for participants of UKB assessment centres of Cardiff, Swansea and Wrexham

Notes: * Values within parenthesis indicate the number of land-use categories within each class.

¹ Ordnance Survey UK (2013) AddressBase Products: User Guide, v1.3 – 06/2013, available at: http://www.ordnancesurvey.co.uk/docs/user-guides/addressbase-products-user-guide.pdf.

² Ordnance Survey UK (2012) AddressBase Products: Classification scheme, v 1.1 – 04/2012, available at: http://www.ordnancesurvey.co.uk/docs/user-guides/addressbase-pluspremium-classification-codes.zip.

³ Department of Transport (2012) NPTG and NaPTAN Schema Guide, 2.4-v0.57, available at: http://www.dft.gov.uk/naptan/schema/schemas.htm#2.4guide.

⁴ The GeoInformation Group Limited (2012) UKMap User Guide. Version 5.

⁵ Ordnance Survey UK (2012) OS MasterMap Integrated Transport Network Layer: User Guide and Technical Specification, v2.0 – 12/2010, available at: http://www.ordnancesurvey.co.uk/docs/technical-specifications/os-mastermap-itn-technical-specification.pdf.

⁶ GeoInformation Group (2012) Image to information: Building class reference sheet, available at: http://www.landmap.ac.uk/images/stories/datasets/building_class100/Building_Class_Reference_Sheet_v6_Sep_12.pdf.

⁷ The Blue Sky colour infrared images and DTMs were procured under an academic licensed through LandMap Services of Manchester Information and Associated Services (MIMAS) at the University of Manchester within the Open Geospatial Consortium (OGC) standards.

Figure 3. Illustration of potential applications of UKBUMP database in BE–health studies.