The United Kingdom National Area Classification of Output Areas

1
The United Kingdom National Area
Classification of Output Areas

• Daniel Vickers
• with Phil Rees Mark Birkin
• School of Geography, University of Leeds

2
PopFest 2004Was held 22nd - 24th June at the
School of Geography, University of Leeds

• Presentations and abstracts can be viewed online
  at
• http//www.geog.leeds.ac.uk/conferences/popfest200
  4/

2
3
What will I be talking about today?

• Introduction to Area Classification and Output
  Areas
• How the Classification system was made including
• What data goes in?
• Methods of standardisation
• Issues of cluster number selection
• Cluster selection
• Cluster Creation
• Naming the clusters
• How well does the classification discriminate
• Census data
• Comparing the Core cities
• Voting patterns
• Deconstructing Rural England
• Mapping the Classification
• Focus on Leeds
• A look around the country

3
4
What is an Area Classification?

• A segmentation system which groups similar
  neighbourhoods into categories, based on the
  characteristics of their residents a
  simplification of complex datasets.

What is an Output Area?

• The smallest area for census output
• 223, 060 in the UK
• EW 174,434 min size 40 hholds 100 people
• Scotland 42,604 min size 20 hholds 50 people
• NI 5,022 min size 40 hholds 100 people

4
5
What Goes In?

• 41 Census Variables covering
• Demographic attributes
• Including - age, ethnicity, country of birth and
  population density
• Household composition
• Including - living arrangements, family type and
  family size.
• Housing characteristics
• Including - tenure , type size, and
  quality/overcrowding
• Socio-economic traits
• Including - education, socio-economic class, car
  ownership commuting and health care.
• Employment attributes
• Including - level of economic activity and
  employment class type.
• How many data inputs are involved?
• 223,060 Output Areas, 41 Variables
• 9,145,460 data points

5
6
The Three Census Problem

• The Census in the UK is run by three separate
  agencies
• ONS in England Wales
• GROS in Scotland
• NISRA in Northern Ireland
• More than just a problem of stitching the tables
  together
• Some tables given different numbers
• Some of the questions on each table are different
• Some of the questions on the tables are in
  different places

6
7
Standardising the Data

• Log Transformation

Why?
Reduces the effect of extreme values (outliers)
Why?
Range standardisation between 0 -1 Problems
will occur if there are differing scales or
magnitudes among the variables. In general,
variables with larger values and greater
variation will have more impact on the final
similarity measure. It is therefore necessary to
make each variable equally represented in the
distance measure by standardising the data.
7
8
Issues of Cluster Number Selection

• When choosing the number of clusters to have in
  the classification there were three main issues
  which need to be considered.
• Issue 1 Analysis of average distance from
  cluster centres for each cluster number option.
  The ideal solution would be the number of
  clusters which gives smallest average distance
  from the cluster centre across all clusters.
• Issue 2 Analysis of cluster size homogeneity for
  each cluster number option. It would be useful,
  where possible, to have clusters of as similar
  size as possible in terms of the number of
  members within each.

8
9
Issues of Cluster Number Selection

• Issue 3 The number of clusters produced should
  be as close to the perceived ideal as possible.
  This means that the number of clusters needs to
  be of a size that is useful for further analysis.
  
• At the highest level of aggregation, the cluster
  groups should be about 6 in number to enable good
  visualisation and these clusters should also be
  given descriptive names.
• At the next level of aggregation, the number of
  groups should be about 20. This would be good for
  conceptual customer profiling.
• At the next level of aggregation, the number of
  groups should be about 50. This can be used for
  market propensity measures from the larger
  commercial surveys.
• (Personal Communication 2003, from Martin
  Callingham, Independent Market Research
  Consultant and Birkbeck College, co-editor of
  Qualitative Market Research Principle and
  Practice, Sage, 2003)

9
10
Cluster Selection

• A three tier hierarchy 7, 21 52 clusters

• First Level target 6, 7 selected based on
  analysis of, average distance from cluster centre
  and size of each cluster.
• Second Level target 20, 21 selected based on
  analysis of, average distance from cluster centre
  and size of each cluster.
• Third Level target 50, 52 selected based on size
  of each cluster. Split into either 2 or 3 groups

10
11
Cluster Creation

• Modified K-means clustering
• First level run as standard k-means
• Second level, first level is split into separate
  files and each file is clustered separately
• Third level, second level is split into separate
  files and each file is clustered separately

11
12
Cluster Creation
12
13
Naming the Clusters
The naming of the clusters is a near impossible
task and one that always provokes much debate.
However, the task is very important, as if it is
done wrongly it can create a false impression of
the people within a cluster. The naming must
follow two general principles 1. Must not
offend residents 2. Must not contradict other
classifications or use already established names.

13
14
How Well Does It Discriminate?
Detached Housing
14
15
How Well Does It Discriminate?
Population Density
15
16
How Well Does It Discriminate?
Indian, Pakistani Bangladeshi
16
17
How Well Does It Discriminate?
Unemployed
17
18
Comparing the Core Cities
18
19
Who do Each Type Vote for?
19
2001 Election Data courtesy of Ed Fieldhouse,
CCSR, University of Manchester
20
Deconstructing Rural England (Devon case study)
Devon Average 31 UK Average 12.5
20
21
Focus On Leeds
Map appears in forthcoming book Twenty-First
Century Leeds Geographies of a Regional City
edited by Rachael Unsworth John Stillwell
Boundaries Community Areas, as defined by Pete
Shepherd, School of Geography, University of
Leeds (built from Output Areas)
21
22
Consultation
62 respondents so far, 33 Academics, 28 Local
Government Two most confused types 4 Blue
Collar Communities 6 Constraints of
Circumstance Easiest type to identify 5 Idyllic
Countryside Consultation to end 4/10/2004
Results as at 10/9/04
22
23
Where would you like to go?
Belfast Brighton Birmingham Bradford Bristol Cambr
idge Cardiff Carlisle Derby Dundee Edinburgh Exete
r Glasgow Hull Ipswich
Leeds Leicester Lincoln Liverpool London Mancheste
r Newcastle Norwich Nottingham Oxford Plymouth She
ffield Southampton Swansea York-
23