The Cartographic Journal Vol. 41 No. 2 pp. 00–00 IGC Special Issue 2004© The British Cartographic Society 2004

Advances in the Human Cartography of the UK

Bethan Thomas and Danny Dorling

Department of Geography, University of Sheffield, Winter Street, Sheffield S10 2TN, UKEmail: danny.dorling@sheffield.ac.uk

INTRODUCTION

This paper concentrates on the mapping of census data andon the UK in particular but hopefully has wider implicationsfor the mapping of human geography more generally bothusing other sources and in other countries. The first detailedcensus atlas of the UK census, excluding Northern Ireland,was drawn by David Rhind and his colleagues using datafrom 1971 and the then revolutionary technology of laserplotting.1 That atlas illustrated over 30 variables in remark-able geographical detail for the time, using the innovativetechnology to individually colour every kilometre square ofthe country according to the social characteristics of thelocal population. In practice the areas the plotter drew wereslightly larger than a kilometre square as they were pushingthe technology of those times to its limits. Thirty years later,conventional maps of the human geography of the UK arevery similar in appearance.

New technology has greatly facilitated the creation ofcensus atlases and has been the key driver in increasingnumbers being produced following each successive censusafter 1971. Following the 1981 census of population manyseparate atlases were drawn of the different countries ofthe UK, and smaller parts of the UK, but none of the UK asa whole and many using what was then called a line plotterto colour areas through ‘hatching’.2 A later publicationexplored the use of mapping with unconventional areas, butchanging the areas being mapped had little impact on thepatterns seen.3 By 1991 technology had advanced but lineplotters were still utilized to produce the first census atlaspublished shortly after the release of that census’s data, thedirect precursor of the maps shown below: People andPlaces 1991.4

Later the Ordnance Survey, then under the direction ofDavid Rhind, produced an atlas based largely on the 1991census. Their atlas used continuous colour to shade areasand was the first of a census in the UK to both be printed infull colour and include Northern Ireland in places, althoughoften as a single spatial unit.5 Two more detailed atlases weredrawn of the same census and of changes since 1981; eachincluded some equal population cartograms, but neitherincluded Northern Ireland and only two-colour separationwas used.6 One used local authority wards as opposed to dis-tricts as its underlying spatial unit, but despite the differencesin scale the resulting patterns were very similar. It is only

recently that production costs have decreased sufficiently toallow a full-colour atlas to be drawn of the whole of the UKwithout the financial sponsorship of a national mapping orcensus agency. The maps shown here are examples from thatnew atlas.7

MAPPING THE 2001 UK CENSUS

Mapping data such as census variables on a conventionalmap of United Kingdom local authorities means that urbanareas with large populations but small areas are virtuallyinvisible to the viewer. Conversely, the large rural areas withsmall populations dominate the map. When mapping dataabout people, it is therefore sensible to use a different spatialmetaphor, one that reflects population size. Therefore weuse cartograms where the area of each local authority isproportional to its population. A population cartogramattempts to maintain topology while showing where peopleare found. Administrative areas with larger populationscover a larger area on the cartogram than those with smallerpopulations. The result is a visualization where the urbanareas where the most people are found are revealed whileareas with lower populations no longer predominate.

For the maps shown here initially a hexagonal cartogramof UK parliamentary constituencies was created. Therationale behind starting with constituencies is that eachconstituency has roughly the same number of electors.Although the ratio of electorate to total population in eachconstituency will differ according to the age structure, thosenot entitled to register to vote and those who fail to register,choosing constituencies gives us the administrative areasthat are most similar in terms of population size. Hexagonswere chosen because an area tends to have just under sixother areas adjacent to it and therefore maintaining topol-ogy is less complicated. However, topological discrepanciesremain; this could have been solved by using a variety ofdifferent shapes but would have made the overall taskmuch more complex and reduced the visual simplicity thathexagons provide.

The next step was to create local authority boundaries.Constituencies do not necessarily fit within local authorities:a local authority may be comprised of several constituencieswhile some individual constituencies may contain two localauthorities and a constituency may fall into two local

I N T E R N A T I O N A L G E O G R A P H I C C O N G R E S S — U K 2 0 0 4 , G L A S G O W

DOI: 10.1179/000870404X00000

CAR50424.pmd 10/18/04, 5:32 PM1

2 The Cartographic Journal

authorities; therefore a best fit had to be applied. Tomaintain the clear visualization, constituencies were onlysplit into at most two parts and then only vertically. Atthe same time the topology of local authorities had to bemaintained. Further geographies such as health authorities,government statistical regions, travel-to-work areas andthe old counties were also determined. This procedure wasa time-consuming manual process. To illustrate why it isworthwhile here we show some of the resulting cartogramsdrawn from the 2001 Census of population, together withthe change since 1991.

FIGURE 1

Figure 1 shows some of the major cities of the UK on botha conventional map and cartogram.

FIGURES 2 AND 3

Figure 2 shows the population and change since 1991 ofeach local authority on a conventional map while Figure 3is of the same data on a cartogram. On the conventional mapit is impossible to discern any detail in London and otherlarge urban areas such as Manchester and Birmingham whilethe more sparsely populated rural areas are visually pre-ponderant. In comparison the cartogram clearly showsthe class of each area and even a single half hexagon, such asthat representing the Isles of Scilly on the tip of the south

west peninsula is easily discernible. The change cartogramclearly shows population growth in the south of the UK anddecreases in much of the north. In the subsequent figuresonly the cartograms are presented.

FIGURE 4

Figure 4 shows population density. Local authorities that aresmall in area tend to have larger populations and thereforehigher population density. London has the highest densityand the Highland Council Area in Scotland the lowest. Asthe local authority boundaries have not changed over thetime period under consideration, change in density since1991 is due to population change. As was shown in Figure 3the population of the UK has moved southwards; as a resultpopulation density in the south has increased, particularlyin London and the South East. Conversely, those localauthorities in the north that have seen declines in thenumber of people living there have also seen a decrease indensity.

FIGURE 5

Figure 5 shows population potential. Population potentialis a measure of how concentrated is the population nearto each area. To calculate it, for each area sum the popula-tion of each other area of the UK, having divided thosepopulations by the distance in metres to the area of interest.

Figure 1. Cities locator map

COLOURFIGURE

CAR50424.pmd 10/18/04, 5:32 PM2

3Advances in the Human Cartography of the UK

Figure 2. Population shown on a conventional map

Figure 3. Population shown on cartogram

COLOURFIGURE

COLOURFIGURE

CAR50424.pmd 10/18/04, 5:33 PM3

4 The Cartographic Journal

Figure 4. Population density

Figure 5. Population potential

COLOURFIGURE

COLOURFIGURE

CAR50424.pmd 10/18/04, 5:33 PM4

5Advances in the Human Cartography of the UK

The populations of geographically close areas have a greatereffect on any particular area than do the numbers of peoplein more distant areas. Population potential is highest inLondon and the South East and lowest in the peripheralareas of Scotland, Northern Ireland the South West. Thechange map shows clearly the gradient of the increase inpotential which declines away from Central London.

FIGURE 6–8

Figure 6 through Figure 8 show the results of modelsbased on census and other variables rather than mappingindividual census variables. The models used for 2001 arein an early stage of development and therefore these arepreliminary mappings.

Figure 6 shows a preliminary mapping of the proportionof households living in poverty in the United Kingdom in2001 and the change since 1991. The poverty measure usedis the Breadline Britain measure (see endnotes) that definesa household as poor if the majority of people in Britain, atthe time of calculation, would conclude that the resourcesavailable to that household constituted living in poverty.Therefore, if living standards rise unequally across society,poverty increases as poorer households are unable to accessthe resources that are seen as necessities. The areas with thehighest rates of poverty are found in East London andGlasgow; the proportion of households living in poverty hasincreased everywhere apart from Northern Ireland.

Figure 7 shows a preliminary mapping of the proportionof households receiving high income and the change since1991. The measure is based on an estimate of those house-holds which have a total net income of over £60,000 a year.Such households predominate in London and the SouthEast, where the jobs that provide such high incomes arefound. The greatest increases are similarly found in London,implying that more high earners are living in the capitalrather than commuting from the surrounding area. The restof the country has generally seen a minuscule increase inthe proportion of high income households, with NorthernIreland having a decrease in such households.

Figure 8 shows a preliminary mapping of the proportionof wealthy households and the change since 1991. Wealth isa measure of assets that a household possesses; for mosthouseholds this is in the form of housing equity. The highestproportion of wealthy households is found in a ring aroundLondon; this is also where the greatest increase has been.These cartograms are a small sample what can be mappedfrom the UK censuses. They clearly show detail that isindiscernible on a conventional map.

CONCLUSION

The first census in the UK to be made available in digitalform was taken in 1971. Although advances in computingwere necessary for this to have occurred, they were not inthemselves sufficient for this innovative step. Instead it was

Figure 6. Percentage of poor households

COLOURFIGURE

CAR50424.pmd 10/18/04, 5:33 PM5

6 The Cartographic Journal

Figure 7. Percentage of households with high income

Figure 8. Percentage of wealthy households

COLOURFIGURE

COLOURFIGURE

CAR50424.pmd 10/18/04, 5:34 PM6

7Advances in the Human Cartography of the UK

the actions of a relatively radical government in the period1964–70 which paved the way both in the appointmentof key civil servants and in the changing of social attitudesto allow for information about the population to be madeavailable to at least a few people in a form which couldbe mapped by computer. ‘People in Britain: A CensusAtlas’ reflected those times and aspirations as much as thetechnology and advances in cartography of the day.

The 1980s and 1990s were much less radical times forthe UK government, the civil service, census taking andthe resulting human cartography. Funding for academicresearch was severely curtailed. Statistical sections of the civilservice were decimated. As computer technology improvedmore atlases were produced but they were often produceddespite the prevailing conditions of their times rather thanbecause of them. A cartographer interested in the newcomputer technology of the late 1970s would probably havehad a far more optimistic view of what the next two decadeswere to bring than what in fact emerged. In UK universitiesvery few appointments were made in the 1980s whilestudent numbers soared and resources were cut. This was aperiod in which academics doggedly carried on, but oftenlooked back to the recent past as a golden era when they hadmore time and the resources for such work.

The late 1990s and the first few years of this centurybrought no new nirvana. The one great advance thatthere has been is that the census results from 2001 are nowavailable to anyone who wishes to study them and eventhe smallest of computers can now handle them. However,almost all mapping software is now designed by large privatecorporations. The vast bulk of maps being produced of the2001 census in the UK are less innovative than those plottedof the census taken over 30 years ago. UK universities, theplaces from which innovations in human cartography of theUK were traditionally most likely to emerge are very differ-ent entities now. There are two million students to be taughtevery year in Higher Education in the UK. Despite this hugenumber of students UK universities have become relianton research contract overheads to fund their operations.Between the increased need for teaching and demand for‘cost recovery’ there is less and less space for innovation.If an innovative new atlas is produced it is now producedwithout funding in the spare time academics have between

conducting the teaching and funded research which they arebeing paid to do.

All these developments are not necessarily to be lamented,but the ivory tower cocoons which have incubated advancesin the cartography of human geography in the UK are lessand less easy places within which to experiment with newideas, methods, technologies and maps. Advances in thehuman cartography of the UK in future may come moreoften from outside the universities and perhaps from outsidethe UK. Indeed it would be refreshing to see the UKmapped from other perspectives. Just as a cartographer inthe 1970s may have been overly optimistic concerning thefuture, we may be a little pessimistic as far as advances incartography in universities are concerned. More positively,new methods of mapping human geography will be devel-oped, but it is becoming increasingly likely that this willhappen outside of the UK, outside of academia, and outsideof the traditional discipline of cartography.

ENDNOTES1 Census Research Unit, Office of Population Censuses and Surveys

and General Registrar Office (Scotland) (1980). People in Britain:A Census Atlas, HMSO, London.

2 Champion, A. (1983). England and Wales ’81, The GeographicalAssociation, Sheffield. ‘Northern Ireland’ was included in Horner,A., Walsh, J. and Harrington, V. (1987). Population in Ireland:A Census Atlas, Department of Geography, University CollegeDublin, Dublin.

3 Champion, A., Green, A., Owen, D., Ellin, D. and Coombes, M.(1987). Changing Places: Britain’s Demographic, Economic andSocial Complexion, Edward Arnold, London.

4 Forrest, R. and Gordon, D. (1993). People and Places: A 1991Census Atlas of England, School for Advanced Urban Studies,Bristol; and Gordon, D. and Forrest, R. (1995). People and Places2: Social and Economic Distinctions in England, School forAdvanced Urban Studies, Bristol.

5 Ordnance Survey (1995). Statlas UK: A Statistical Atlas of theUnited Kingdom, Ordnance Survey, Southampton.

6 Dorling, D. (1995). A New Social Atlas of Britain, John Wiley andSons, London; and Champion, T., Wong, C., Rooke, A., Dorling,D., Coombes, M. and Brunsdon, B. (1996). The Populationof Britain in the 1990s: A Social and Economic Atlas, OxfordUniversity Press, Oxford.

7 Dorling, D. and Thomas, B. (2004). People and Places: A 2001Census atlas of the UK, The Policy Press, Bristol.

CAR50424.pmd 10/18/04, 5:34 PM7