For the purposes of the SLOPE project, we have collated and constructed a variety of geographical indicators. These geographical indicators are included as potential predictors in our modelling of childhood obesity. We have collated these data at the Lower and Middle layer Super Output Area ( LSOA/MSOA ) levels. These are geographical units with average populations of 1,500 and 7,000 respectively, and an average size of 4km 2 and 21km 2 . This page will outline the extent and sources of data we have collated so far.
Income Support data
In the UK, Income Support is a means-tested benefit for individuals or families on a low income. We included the proportion of the local population claiming income support. These data are available quarterly, and were taken from https://www.nomisweb.co.uk/datasets/bissa .
Socially rented households
In the UK, households that are rented from a council or housing association are colloquially referred to as socially rented. Rental rates for social housing are typically cheaper than those from private landlords, and are awarded based on a needs-based priority system . We included the proportion of households that are socially rented from the 2001 and 2011 Censuses. 2001 data are linked to children born 2003-2006, 2011 data are linked to children born 2007-2018. These data are available at https://www.nomisweb.co.uk/census/2001/ks402ew and https://www.nomisweb.co.uk/census/2011/ks402ew .
Indices of Deprivation
The indices of deprivation are a measure of how deprived the local population are, relative to the rest of England and Wales. The indices of deprivation are often referred to as the ‘Index of Multiple Deprivation’. We collated data from the 2004 , 2007 , 2010 and 2015 indices of deprivation. The indices of deprivation were calculated at the MSOA level using the methodology outlined in Appendix A of the 2015 report. Children born 2003-2005 were ascribed data from the 2004 indices, children born 2006-2008 were ascribed data from the 2007 indices, children born 2009-2012 were ascribed data from the 2010 indices, and children born 2013-2018 were ascribed data from the 2015 indices. The indices are standardised to England and Wales, and we also calculated relative deprivation for the county of Hampshire.
House prices
Records of all household sales in England and Wales are accessible from the Land Registry . We extracted all sale location and price data for all flats and houses sold 2003-2017 inclusive. We then calculated the annual median price for these sales, and converted these into quintiles in comparison to the remainder of England and Hampshire.
Air pollution
Average levels of background air pollution are modelled on an annual basis by the Department for Environment Food and Rural Affairs . Several metrics were extracted, including PM 2.5 , PM 10 and NOx. Annual means for each metric are provided for 1km grids across the UK. We calculated a spatially-weighted average for each LSOA and MSOA.
We generated a number of area-level indices from Ordnance Survey and Open Street Map data which we cannot share due to the agreement terms with Ordnance Survey.
Green space
Green space coverage (natural land) was measured using annual releases of the Ordnance Survey’s MasterMap Topography Layer (2006-2016). Green space areas were identified as polygons tagged with ‘natural environment’, with the exception of freshwater areas and marshes. The area of these polygons were summed to calculate the proportion of the area covered by green space. Births in the years 2003-2005 were ascribed the value for 2006.
Supermarket density
Supermarket density was calculated from annual releases of the Ordnance Survey’s Points of Interest dataset (2007-2017). Kernel-density estimation was used. This technique estimates the concentration of points with a spatial decay parameter, such that further points are given a lower density value. Kernel-density estimates were generated for a radius of 800 metres around each supermarket in each year at 100-metre intervals, because we believed this to be a reasonable walking distance. These estimates were then averaged across each LSOA and MSOA. Children born before 2007 were ascribed the value for 2007.
Unhealthy food index
An unhealthy food index was generated from annual releases of the Ordnance Survey’s Points of Interest dataset (2007-2017). As above, kernel-density estimates were generated separately each year for unhealthy (bakeries, confectioners, convenience stores and takeaways) and healthy (grocers, markets and health food stores) food outlets. The densities for both types of outlets are averaged across the area, any the relative exposure to unhealthy outlets is calculated by subtracting the density of healthy outlets from that of unhealthy outlets, in line with a previous study. Positive values indicate a greater exposure to unhealthy outlets, and vice versa. Children born before 2007 were ascribed the value for 2007.
Spaces for social interaction
Spaces for social interaction were extracted from annual releases of the Ordnance Survey’s Points of Interest dataset (2007-2017) and the UK Government’s ' get information about schools ’ service. We defined spaces for social interaction as places that encourage families with young children to visit, which involve social and physical activities. The proximity of these destinations likely encourage their use, and the likelihood of families walking to them, thus promoting physical activity. Our selection was informed by recommendations on the Mumsnet internet forums for places that offer classes for young children, or public spaces where young families could meet. From the Points of Interest data, cafes, community centres, gymnasiums/leisure centres, libraries, places of worship, playgrounds, soft play centres and swimming pools were selected. Data on opening and closing dates for children’s centres (sometimes referred to as ‘Sure Start centres’) were derived from the UK Government’s ‘ get information about schools ’ service. Similar to the food indices, kernel-density estimates (with an 800 metre radius) were created for each year for the total sum of spaces for social interaction, and averaged across the area. Births before 2008 were assigned the value for 2007.
Walkability
A walkability index was generated, with residential density, gradient change, intersection density and land-use mix as its components. Residential density is the number of households in the area in the 2011 Census , per km 2 . Gradient change is the average slope (in degrees) between 5-metre intervals, derived from Ordnance Survey’s OS Terrain 5 . Intersection density is the number of road junctions with 3 or more exits per km2, derived from the Ordnance Survey’s ITN Layer. Land-use mix is an entropy score, where 0 indicates that the area is covered by land-use type exclusively, and 1 indicates that the area is perfectly split between all land-use types. The land-use types include: residential, retail, office, entertainment, institutional and greenspace. The overall index is a sum of z-scores for each indicator, relative to the county of Hampshire, with the intersection density component being weighted twice. The walkability index was calculated using the latest available data, not on an annual basis.
The University cannot accept responsibility for external websites.