GIS and census design - Amazon S3 · Census Programme GIS and census design Professor David Martin School of Geography, University of Southampton. ... Automated GIS …

Census

Programme

GIS and census design

Professor David Martin

School of Geography, University

of Southampton

Census

Programme

GIS and census design: what’s the connection?

Census

Programme

What’s the connection?

• Census as driver for

neighbourhood geographies

• GIS for census geography design

• Neighbourhood definitions

prominent in public policy

Census

Programme

Overview

• UK population census basics

• 2001 census geography

• Neighbourhood statistics

• Towards 2011

– Address listing

Census

Programme

UK population census: basics

• 10-yearly data collection

• Questionnaire delivered

to every household

• Legal obligation: 96%

coverage

• Multiple uses and users

• Single most important

social data source for

small areas

Source: ons.gov.uk

Census

Programme

Census topics

• Age, sex, marital status,

household structure

• Country of birth/ethnicity

• Health/disability

• Economic status

• Housing, amenities

• Travel to work and cars

• Qualifications

Source: ons.gov.uk

Census

Programme

100+ years of census mapping

Charles Booth: poverty

in Pimlico 1898-9

Neighbourhood

Statistics: households lacking amenities 2001

Source: neighbourhood.statistics.gov.uk

Source: booth.lse.ac.uk

Census

Programme

Need for information

• Policy Action Team 18:

Neighbourhood policy requires

accessible baseline data

• No definitive neighbourhoods

• Definitional anarchy

– Neighbourhood,

– community, ward,

– catchment…

Census

Programme

2001 census geography

Census

Programme

Previous census geographies

• Enumeration-based EDs

• Wide variation in population,

some sub-threshold: no data

• Significant manual redesign at

each census

• No match to postcode geography

• Not ‘neighbourhood’ building

blocks

Census

Programme

GIS for enumeration design

Source: ONS

Census

Programme

Case for redesign

• Demands from census users:

– Postcode and statutory geographies

– Uniformity of population sizes (all above

threshold)

– Control over shape (observe settlement

pattern and topographic features)

– Internal homogeneity of population

– Compatibility with previous census

geographies!

Census

Programme

Overview of design methods

• Synthetic unit postcode building block

polygons

– Nesting within wards (Dec 2002) and

parishes, incorporating topographic

features

• Automated zoning procedure

– Iteratively recombination of building

blocks seeking best trade-off of design

constraints

Census

Programme

Address-based Thiessen polygons

• Thiessen polygons

around individual

ADDRESS-POINTS

intersected with ED,

ward, parish

boundaries and road

centrelines

Census

Programme

Unit postcode building blocks

• Address polygon

boundaries dissolved

to form unit postcode

polygon building

blocks

Census

Programme

AZP (Openshaw, 1977)

• Initial random aggregation of building

blocks into output areas within ward

or parish polygon

• Compute design statistics

• Evaluate effect of possible swaps of

building blocks between output areas

• Keep swaps that improve the overall

solution

Census

Programme

Automated output area design

Group addresses into

postcode building blocks

Group addresses into

postcode building blocks

Automated GIS-based

output area design

procedure

Automated GIS-based

output area design

procedure

Design Constraints

(Contiguity, Thresholds,

Shape, Size, Homogeneity)

Design Constraints

(Contiguity, Thresholds,

Shape, Size, Homogeneity)

2001 Output

Areas

2001 Output

Areas

Census

Programme

OA design (1)

Initial random

aggregation of

postcodes into

potential output

areas

Census

Programme

OA design (2)

Choose one

postcode at

random as

candidate for

swapping into a

different output

area

Census

Programme

OA design (3)

Make the swap

and evaluate the

impact on the

overall solution

Census

Programme

OA design (4)

If swap does not

result in an

improvement, go

back to the

previous

configuration

Census

Programme

OA design (5)

Choose another

postcode at

random as

candidate for

swapping into

another output

area

Census

Programme

OA design (6)

If the swap results

in an overall

improvement,

keep it as part of

the solution and

examine a new

potential swap…

Census

Programme

Contiguities, thresholds, urban/rural

• Output areas assembled from contiguous postcodes (NB treatment of stacked postcodes)

• Output areas above 100 person and 40 household thresholds (NB treatment of sub-threshold parishes)

• Initial postcode classification to urban/rural based on DETR boundaries

Census

Programme

Size and shape

• Output areas should be as

uniformly sized as possible; target

125 households

minimize Σ(OApop-target)2

• Output areas should be as compact as possible

minimize distance to OA centroid…

Census

Programme

Distance to OA centroid

PC mean

centroid

OA centroid

Census

Programme

Intra area correlations

• Maximize intra-area correlations

(IAC): ratio of area level to individual

level variance

• Higher correlations ~ greater internal

homogeneity

• Tenure (4) and dwelling type (7)

categories used

Census

Programme

Combination of constraints

• All constraint statistics recomputed at

each iteration

• Must always meet contiguity and

threshold requirements; urban/rural if

possible above threshold

• Population, shape and homogeneity

constraints combined with equal

weighting

Census

Programme

Digital census geography

• 2001 Output

Areas (England

and Wales)

• Mean 297

persons; 123

households

Census

Programme

UK geography?

• Northern Ireland: same algorithm

as England and Wales but

digitized postcode boundaries

• Scotland: own algorithm to

maximise continuity with 1991

output areas

Census

Programme

Neighbourhood statistics

Census

Programme

Neighbourhood statistics

• NeSS as primary delivery

mechansim – originally 1998 wards

• Addition of census, and

administrative data streams

• Post-2001 geography built on

postcode > census output area >

Super OA

• Cannot continually reconstruct

small areas: data confidentiality

Census

Programme

NeSS geography

• Aggregated source records from

multiple suppliers

• Single hierarchy based on census

OAs (and hence primarily on

postcodes, with lookup tables)

• Estimation of attribute values to

standard geography if necessary

• Intended as durable core geography

Census

Programme

• OAs as building blocks of NeSS

geography

• “Lower” (pop mean 1500; n=35000)

– Automated zone design, nesting within

census wards: now created

• “Middle” (pop mean 7500; n=7000)

– Automated design, not constrained to

wards, some local consultation

Super Output Areas

Census

Programme

Key

Lower Layer Zones

Greenwich lower layer SOAs

Census

Programme

New SOA data streams

• Indices of Deprivation 2004,

2007

• Fires attended

• National curriculum assessments

• Benefits claimants

• Land use

• Air quality

Census

Programme

Census

Programme

Census

Programme

Census

Programme

Neighbourhoods?

• More flexible building blocks and richer data than previously

• Irregular boundaries

• Little regard to landscape features

• Little regard to social geographies

• Never intended as definitive ‘neighbourhoods’

Census

Programme

Towards 2011

Census

Programme

Towards 2011

• Essentially conventional 2011

census

• Mail-out and mail-back

• Critical nature of address lists

• Internet data collection

• 4-page per person questionnaire

• Big focus on population count

Census

Programme

Census

Programme

Complex geography requirements

• Stability and change

• Data supply

• Landscape features

• Social homogeneity

• Reflect contemporary housing

development

• ‘Real’ neighbourhoods!

Census

Programme

Address listing

• No definitive address list in

England and Wales

• USA, Scandinavia, South

Africa… seen as essential

• Address Point/Postcode Address

File/Mastermap address layer

• National Land and Property

Gazetteer

Census

Programme

Address strategy

• Use combination of OS

Mastermap AL 2 and NLPG

• Pre-listing and checking by

enumerators

• Can the information be shared?

• Counting the Population review –

possible action on addressing?

Census

Programme

Where do we go from here?

• Proposed principle of stability

• Research on residential change

– based on address and

postcode geographies

• Evaluation of population change

requiring essential OA splits and

mergers <5%

• Need maintenance methodology

• Need disclosure policy details

Census

Programme

UK geography?

• Northern Ireland: total change -

enforced

• Scotland: minimal change –

consulting now

• Convergence to whole-UK

approach extremely unlikely

• Counting the Population –

definitive address list???

Census

Programme

Research…

• Methods for improving OA

boundaries

• Data ownership issues more

important barriers than technical

ones

• How best to reflect housing

change?

• Small area geography beyond

census?