Category Archives: mapping

The geography of social mobility

Yesterday’s DfE report Unlocking Talent, Fulfilling Potential: a plan for improving social mobility through education put a strong emphasis on ‘place’, recognising that how young people ‘get on’ in life cannot easily be separated from the areas in which they grow up. I’ve blogged before, here and here, on the strong links between young people’s outcomes and the neighbourhoods they grow up in. Here, we’ve conducted new analysis on fresh data from UCAS to move the debate further.

‘Place’ and social mobility
Justine Greening promised that this week’s new social mobility strategy would be responsive to the context of local communities, and this is welcome. But aspects of ‘place’ – such as young people often being fiercely attached to their neighbourhood – make simplistic narratives of ‘social mobility’, and their expectation that young people can easily uproot themselves to capitalise on opportunities, quite problematic. I’ve argued so before here.

Even so, if ‘place’ is to be at the heart of our new social mobility strategy, we can do worse than try to develop a nuanced understanding of the types of places where young people’s outcomes fall short of their peers. This week UCAS released data from its 2017 End of Cycle Report which shows the entry rate to higher education (HE) of 18 year olds by parliamentary constituency. Constituencies cover mid-sized areas (compared to the small units of the POLAR classification and larger units like regions), and this makes them a handy geography for picking out geographical ‘hot spots’ and ‘cold spots’ in young people’s outcomes. As we’ve argued elsewhere, constituency-level data are also a useful tool when trying to mobilise political will behind a policy problem.

What do the data show?
Looking across all constituencies, there is a wide spread of HE entry rates, from the lowest (Glasgow North East, 19%) to the highest (Wimbledon, 81%). The largest increases in entry rates over the last decade appear to be in core metropolitan areas like Glasgow, Leeds, Manchester and Birmingham. In the capital, East London appears to have experienced the largest increase in entry rates.

Full map here

Beyond picking out the entry rates in individual constituencies, one way of looking for broader patterns is to categorise constituencies into ‘types’. When we do this based on the urban/rural mix in each constituency, we see that a particular set of constituencies (classified as ‘urban with minor conurbations’) have the lowest entry rate, on average – 35% compared to 42% across all constituencies.

If we give these ‘low participation’ constituencies a black border we can see they lie in a tightly focused cluster between Derby and Nottingham, and between Sheffield, Barnsley and Doncaster.

Full map here

These aren’t necessarily the individual constituencies with the lowest entry rates – they’re shown in red on the map, with notable clusters of low participation in the Thames Estuary, East Anglia, the South Wales Valleys and the Central Belt of Scotland. However, as far as there are types of constituency that seem to send relatively few young people on to higher education, there appears to be a distinct cluster of former mining areas in the centre of England.

Why might these areas be significant?
Young people growing up in inner-urban areas of large cities tend to face the highest levels of deprivation. However, inner-urban areas also give young people access to a range of amenities, host a wide mix of people and jobs, and have historically tended to receive the most policy focus – from regeneration programmes to teacher recruitment schemes. It appears that young people might face particular challenges growing up in smaller urban areas: the outskirts of cities, smaller towns, or the places where one conurbation ends and another begins. It might therefore be worth policy makers, Widening Participation initiatives and, perhaps, social mobility strategies, focusing more on these types of area in the future.

I originally wrote this blog post for LKMco.

Mapping school exclusions

5170 pupils were excluded from state-funded primary, secondary and special schools in England in 2011/12. Since peaking in 2003/4, numbers of school exclusions have been falling for the last decade, although other trends have been more persistent: the vast majority of permanently excluded pupils are boys, and Black Caribbean pupils are around three times more likely to be excluded than the school population as a whole. A 2007 report estimated the cost of each exclusion, to the individual pupil and to society, as just under £64,000.

Using the latest data from the DfE, I’ve mapped permanent exclusions in England to see what geographical trends exist.

At a local authority level, the highest rates of exclusion are in North East Lincolnshire, Reading, Central Bedfordshire and Bury. At a regional level, the highest rates of exclusion are in London and the Midlands, with the lowest rate in Yorkshire and the Humber. Interestingly, there’s no correlation between an authority’s exclusion rate and its GCSE performance.

Poor outcomes in affluent areas

Ofsted released its annual report today, and one of its key messages is that some of the poorest educational outcomes are now concentrated in more affluent outer-urban areas. Traditionally, the inner city has been the place to go if you want to find the weakest GCSE results, the highest rates of school absences and the lowest levels of transition on to further and higher education. Since the success of urban-focused policies such as the Manchester and London Challenges, however, area-based patterns in educational outcomes have started to shift, and the link between deprivation and educational outcomes shows signs of weakening.

The 2010 Index of Multiple Deprivation offers a simple way of illustrating this trend. The IMD ranks areas based on their overall level of deprivation, and also according to the educational outcomes of the young people who live there (the Children and Young People sub-domain). Comparing ranks on these two measures, it’s possible to identify two types of area: deprived areas with relatively strong educational outcomes, and less deprived areas with relatively weak educational outcomes. I’ve mapped the data for London, and the pattern is interesting. Deprived parts of inner and east London appear to have relatively strong educational outcomes, whereas many of the capital’s more affluent suburbs have relatively weak educational outcomes. It’s not a cast-iron trend, but it does seem to lend some support to Ofsted’s claims, and suggests that we need to have a more nuanced discussion about the types of area where young people’s outcomes are poorest. It may be that deprivation is no longer the most powerful compass.

Fullscreen map here.

Counter-deprivational outcomes

The Index of Multiple Deprivation (IMD) is one of the core area-level measures in my research, which is looking at how young people’s occupational aspirations are shaped by the areas they live in. IMD scores are a composite of several specific deprivation ‘domains’, which separately measure a range of types of deprivation, from an area’s income deprivation and health deprivation to its education and skills deprivation. Lots of research and policy is based on analysis which links outcomes with area-level deprivation, based on overall deprivation scores. Indeed, current government aspirations policy is based on analysis which contends that young people’s aspirations are lower in more deprived areas. But much of this analysis, based on overall deprivation scores, overlooks two important points about the way deprivation is measured by the IMD:

  1. A high level of overall deprivation doesn’t necessarily mean an area is highly deprived on all deprivation domains
  2. Area-level outcomes may be due to particular forms of deprivation rather than overall deprivation

This throws up two interesting types of target area for deprivation-based research like mine: areas that have high overall deprivation scores, but low scores for specific deprivation domains; and areas that have have low overall deprivation scores, but high scores for specific deprivation domains.

I’m interested in one deprivation domain in particular – the Children and Young People domain. This deprivation domain measures an area’s Key Stage 2, 3 and 4 results, progression rates to further and higher education, and school absences. Essentially, then, it represents area-level educational deprivation. So far my analysis suggests that aspirations don’t seem to be lower in more deprived areas, when measured by their overall IMD score. Young people in the most deprived areas seems to have aspirations that are just as high as those of their peers from less deprived areas, and this finding runs contrary to the thinking behind current government policy. However, it does seem that educational deprivation is linked with lower aspirations.

So, I was interested to see whether, and where, the following two types of area existed:

  1. Areas with high levels of overall deprivation but low levels of educational deprivation
  2. Areas with low levels of overall deprivation but high levels of educational deprivation

I decided to focus on London because it’s a large city with a wide range of area types, including some of the most and least deprived areas in the country. The map below shows the geographical distribution of these two area types:

The north and east of the city contain many deprived areas which nonetheless have low levels of educational deprivation. Meanwhile, the more affluent south west and south east, along with a number of more peripheral suburban areas, contain areas with low levels of deprivation but relatively high levels of educational deprivation. The map demonstrates the existence of deprived areas with relatively strong educational outcomes (and, based on the initial findings from my research, probably sites of high aspiration), alongside less deprived areas with relatively weak educational outcomes (and therefore probably sites of  low aspiration).

I think of these areas as exhibiting counter-deprivational outcomes. They’re particularly interesting sites for research, but they’re all too easy to overlook.

Mapping the housing shortfall

There’s been a lot of discussion today about the housing shortfall in the UK, and in particular whether the restrictions of the planning system or the cap on councils’ housing spending is to blame. Whatever lies behind the low present rate of house building – and there are a range of possible explanations, including the shape of the mortgage market and the economy more broadly – it’s clear that the rate at which we’re building new homes in England has some way to go until it returns to pre-crash levels:


It’s difficult to calculate exactly how many new houses we need to build. Firstly, on the supply side, homes can be created by renovating existing, empty properties – not just by building new ones. Secondly, on the demand side, housing need isn’t just embodied by those who are homeless but by those living in overcrowded homes, temporary accommodation or in housing which is unfit for habitation. While in the private market, housing need doesn’t always become apparent until individuals or households decide to try and buy or rent a new property, in the social rented sector there is a definite number attached to the shortfall between housing demand and supply: the housing waiting list.

The map below shows the extent of the housing waiting list in each of the 364 local authorities in England, according to the latest data which dates from April 2012.

Unsurprisingly there’s a lot of variation, with waiting lists ranging from below 100 to the tens of thousands. This is because the numbers are counts rather than proportions, and some authorities have much larger populations than others. However, there’s a power to keeping the data in this form, as raw numbers: if you’ve never looked at the housing waiting list before it’s quite staggering just how large it is, and how quickly it’s growing. Ten years ago there were just under 1.1 million people in the queue; today that figure has risen to over 1.8 million. In metropolitan authorities such as Sheffield, where the housing waiting list is over 68,000, the extent of housing need is astounding.

Suspended spaces

Last month I contributed a post to the cities@manchester blog which explored the ‘suspended spaces’ that exist amidst the urban bustle of Manchester city centre. During my four years living in Manchester I became increasingly interested in – and often frustrated by – the temporary car parks, abandoned buildings, stalled construction projects and boarded-up voids that pepper an otherwise growing, resurgent cityscape. These suspended spaces, as I decided to call them, have a variety of origins – from the vagaries of the post-recession housing and credit markets to the never-realised grand plans of the city council. They are sometimes host to innovative and unexpected uses, providing a free backdrop for bandshoots or a place to store stuffed animals. Some suspended spaces are well known – indeed unmissable – such as the giant crater of Origin, while some are a little more off the beaten track, such as the Kensington gap (and its resident blue tit).

My coverage of the suspended spaces in Manchester was by no means exhaustive and I invited people to contribute to my map, which is open for collaboration by anyone with a Google account. At first: radio silence. Then: a welcome update from Parkstarter, a Manchester-based initiative to reclaim poorly used urban spaces in the city. They recently transformed the temporary car park on the former Dobbins site on Oldham Street into a day-long popup park – complete with real grass. A fitting example of the opportunities presented by suspended spaces, and the way in which these are often missed.

Priced out

Housing Benefit was introduced in 1988 as a subsidy to both social-rented and private-rented tenants as a way of increasing access to affordable housing for those on low or no incomes. HB arose in response to the shrinkage of the social housing sector caused by a reduction in new social housing construction and the sale of social houses through the Right to Buy. Since its introduction, the government bill for HB has increased from £2.5 billion in 2001 to £20 billion in 2010, leading to efforts by successive governments to bring HB spending down. In April 2011 the coalition government introduced a number of reforms to HB with the aim of saving £1.8bn in expenditure by 2014/15. The two main measures in the package of reforms were a reduction in the basis for setting the Local Housing Allowance from the median of local market rents to the 30th percentile of local market rents, and the introduction of nationwide caps on LHA, with specific caps for each size of dwelling.

The Local Housing Allowance is used to calculate how much Housing Benefit someone can claim. LHA rates vary across the country according the level of local market rents, and are calculated by Broad Rental Market Areas. Before April 2011, LHA was set at the median of market rents in a given BRMA; since April 2011 LHA is set at the 30th percentile. Setting LHA in accordance with the 30th percentile of market rents as opposed to their median means that, in theory, in any given local housing market, an individual claiming Housing Benefit is restricted to the bottom third of the rental market, as opposed to the bottom half (unless they’re able to top up from their own funds). In real terms, the recalculation of LHA based on the 30th percentile of local rents translates into an average loss of £9 a week for someone in a 2-bed property, £12 a week for someone in a 3-bed property and £18 a week for someone in a 4-bed property. These losses are substantially greater in London, where market rents are higher across the board: the recalculation of LHA equates to weekly losses of £37, £56, £72 for 2-bed, 3-bed and 4-bed properties respectively in Islington and £58, £91 and £161 in Kensington and Chelsea.

The 2011 reforms also introduced a nationwide cap on LHA rates of £250 a week for 1-bed properties, £290 a week for 2-bed properties, £340 a week for 3-bed properties and £400 a week for properties with four or more bedrooms. In BRMAs with market rents below the national average these caps are less significant, as the level of LHA in these areas will likely come out below the level of the cap for each size of dwelling. However, in places like London with high market rents the LHA caps are significant. Indeed, capping the level of LHA was specifically intended to deal with the problem of very high Housing Benefit claims in places like inner London, and this much becomes clear when we map the data.

Since April 2011 LHA data from the Valuation Office Agency have only been available in capped form, so it’s possible to see where 30th percentile market rents are equal to or greater than the new capped LHA rates, but it’s not possible to see by how much. In other words, we can see which areas are subject to the cap but not to what extent they’re subject to the cap. This data is shown in the map below, and what’s clear is that the impact of the new LHA caps is currently only felt in London. In theory, in those parts of London shaded red below, the new LHA caps restrict people on Housing Benefit to below the bottom third of the private rented housing market. Here’s the map in fullscreen.

Using slightly older data from 2011, just before the introduction of the new LHA regime, we can get a better idea of the extent to which different areas of London are affected by the new caps. The map below shows, for London BRMAs, the amount by which the January 2011 30th percentile market rent for a 3-bed property falls above (red) or below (green) the new LHA cap of £340 a week. In Inner West London, people on Housing Benefit face a shortfall of £5 a week if they want to live outside the bottom third of the housing market and  aren’t able to top up with their own resources. In Central London the shortfall is over £300 a week. Here’s the map in fullscreen.

These maps show that people on Housing Benefit are being ‘priced out’ of large areas of central and inner London, and in many cases the stock of property in inner London that remains affordable to those on Housing Benefit is restricted to well below the bottom third of the market. Data from the Department for Work and Pensions (Table 9) show that only 7% of private rented accommodation in central London is now affordable for those on Housing Benefit, as opposed to 52% before the 2011 reforms.

Where is the Olympic Torch?


The burning question. You can see the route of the Olympic Torch relay here on the official website for the London 2012 Games, but I wanted to make my own map. The route of the Olympic Torch relay is published by LOCOG and available as a file here, courtesy of Oliver O’Brien at the Centre for Advanced Spatial Analysis, who’s also made his own map of the torch relay route. The data is in JSON (JavaScript Object Notation) format, so needs to be converted before Excel or Fusion Tables will be able to read it. Fortunately there’s a JSON to CSV converter online. The file has longitude and latitude data for each calling point on the route, so setting up markers for each stop on the relay is easy. The tricky part is plotting the route itself. A Fusion Tables map can draw a line between two points, but this means creating a new column with a multigeometry element, which is quite easy to do manually if you only have a few points to connect together but very time consuming if you have over 1000. Here’s an Excel macro I wrote to do the job automatically. You just need a column with latitude data and a column with longitude data, with data in rows so each row represents a point defined by a latitude and longitude. The latitude column needs to be to the left of the longitude column. Create a blank column to the left of both, and run the macro in any empty cell in the blank column. It’ll create a multigeometry element with a line connecting the location of the current row with the location in the row directly above. 1000 clicks later, and you have your very own Olympic Torch route map.

Mapping deprivation in Manchester

Here’s a map of deprivation in Greater Manchester. The most deprived areas are shaded dark red; the least deprived areas dark blue. Each small region on the map is a Lower Super Output Area (LSOA). There are 34,378 LSOAs in England and Wales, and on average an LSOA covers an area with a population of just 1500 people, so data at LSOA level is very high resolution. This is good for mapping deprivation, which can vary suddenly and subtly between areas – as is clear in the case of Manchester. Particularly interesting are places like Wythenshawe: deprived areas in a sea of relatively less deprived areas. Things are best viewed in fullscreen.

Mapping with Fusion Tables

This is my first stab at using Google Fusion Tables to present area-level data. In this case, it’s just a simple map of the Greater London Authority constituencies, coloured according to the results of the 2012 GLA elections. Fusion Tables are one of the easiest ways to host and publish data online – especially maps.

The boundary data comes as a shapefile from UKBORDERS. It’s then a case of converting the shapefile into a set of KML files – one KML for each constituency – which can be imported into individual rows within a Fusion Table, along with attributes contained within the shapefile’s related dbf file, like constituency name. All this can be done using the web-based tool shpescape. Then it’s just a case of adding data to the Fusion Table, such as the election result for each constituency, along with some adjustments to the map style to display each constituency in a certain colour according to the winning party.

This is quite a boring map. Ultimately I want to map more interesting things – ideally, things other people haven’t mapped yet.