Spatial Geography

1

• Spatial Geography

2
Objectives

• Why Spatial?
• ASGC
• Elements of ASGC
• Spatial Accuracy
• Temporal Comparisons
• MAUP
• Social Atlas

3
Why Use Space?

• Why do we want to use space?
• What are the options?
• Tables
• Graphs
• Maps

4
Using Census OutputTable
5
Using Census DataGraph
6
Using Census Data Map
7
Using Census Data

• Which of those is most informative?
• Spatial is complete package both data and the
  distribution.
• Move onto the Spatial context for ABS data.

8
Spatial Geography

• These data need a spatial basis for analysis.
• The ABS provide the Australian Standard
  Geographic Classification.
• This is a hierarchical spatial classification
  which starts with the collector district and by
  aggregating culminates in SLAs,SSDs, SDs, States
  and Australia.

9
AUSTRALIAN STANDARD GEOGRAPHIC CLASSIFICATION

• Basis for all ABS data
• The spatial units are hierarchical (State -
  Statistical Division-Statistical - Statistical
  Local Area - Collector District)
• Data can be aggregated or disaggregated
• provide a unit which has some meaning in relation
  to social or economic linkages
• CD is the smallest spatial unit in the Australian
  Standard Geographic Classification (ASGC)

10
ASGC
Source ABS 1999, Australian Standard Geographic
Classification, Catalogue No 1216
11
Australian Statistical Geography
12
ASGC

• Many spatial options.
• Focus on the CD as this is the smallest unit
  available.

13
CD DESIGN

• A ten day field workload for a census collector
• In urban areas CDs average about 300 dwellings
• In rural areas the number of dwellings per CD
  reduces as the population density decreases
• CDs in aggregate cover the whole of Australia
  without gaps or overlaps
• CD should follow natural boundaries where
  possible (eg road centrelines, rivers)
• Must not cross SLA or other ASGC boundaries
• Sensible and logical and change as little as
  possible
• Protect confidentiality

14
AUSTRALIAN STANDARD GEOGRAPHIC CLASSIFICATION -
USER REQUIREMENTS

• Flexible enough to allow statistics to be
  produced for a variety of different
  customer-specified spatial units
• stable over time (especially in the case of CDs
  and statistical local areas ) and
• Improved, making them more homogeneous in terms
  of land use or urban/rural characteristics.

15
Collection Districts

• In metropolitan areas CDs average approximately
  220 dwellings.
• In rural areas the number of dwellings per CD
  reduce with lower population densities.
• In 1996 34,500 CDs in Australia.
• In 2001

Source ABS 1999, Australian Standard Geographic
Classification, Catalogue No 1216
16
Collection Districts

• As the ONLY available small area unit it has
  become an analysis unit it this appropriate?
• Why?
• Issues with CD.

17
CD CHANGE

• Before each census CD boundaries are reviewed
• CDs may be changed due to collector workload
• Changes to legal boundaries such as LGAs
• Growth (as the number of dwellings increases the
  CD is split)
• Decline (may result in merged CDs)
• Re-alignment to existing administrative
  boundaries (eg postcodes or suburbs)

18
Comparability codes for the 1991 - 1986 census

• Comparability code Description
• 0 1991 CD is perfectly comparable to the 1986 CD
• 1 1991 CD is comparable - boundary change but
  population dwellings unchanged
• 2 1991 CD is comparable - within a 2 dwelling
  limit
• 3 1991 CD is comparable within a 10 dwelling
  limit
• 4 1991 CD is not directly comparable- 1986 CD has
  been split into two parts, with the 1986
  boundary being retained around the two new 1991
  CDs
• 5 1991 CD is not directly comparable.-1986 CD has
  been split into three or more, with the 1986
  boundary being retained around the three of more
  new 1991 CDs
• 6 1991 CD is not directly comparable due to a
  split or boundary variation with no common
  boundaries being retained
• 7 1991 CD is not directly comparable due to an
  amalgamation of two 1986 CDs to give a new 1991
  CD ( the 1986 outer boundaries are retained)
• 8 1991 CD is not directly comparable due to an
  amalgamation of three or more 1986 CDs to give a
  new 1991 CD ( the 1986 outer boundaries are
  retained)

19
Numbers and types of changes to Census Collection
District boundaries between the 1991 - 1986 census
20
Temporal Comparison

• OK if no changes
• Difficult if many changes.

21
Spatial Accuracy

• Accuracy in space
• Precision
• Point in time focus
• Change in CDs over time poses a series of
  problems.

22
1996 CD
23
Spatial Accuracy
50 m
62.8 m
1991 CD
1996 CD
24
CD Change
1991 CD
1996 CD
25
Concordances

• Highlighted a number of problems.
• Solutions.
• Concordances.
• Enable conversions between data from different
  spatial units.
• Enable of time series analysis where the spatial
  unit changes over time
• Boundary changes and
• When using non standard units.

26
Solutions

• Two solutions

CD Match
Build comparable units over time
Concordance
Spatial Concordances
27
PROCESS FLOW CHART
CREATING MATCHED CDs
1991 Census Boundaries
28
Dwelling Based Concordances

• Use actual dwelling data to determine pattern of
  spatial unit.
• Design new units and need to allocate population
  data to new units.
• Pro rata on distribution of dwellings in original
  unit and new unit.

29
(No Transcript)
30
Modifiable Areal Unit Problem

• Geographers intrinsically analyse data based upon
  areal units.
• Little attention has been focused upon the
  definition of spatial units.
• Small spatial units are aggregated to provide
  analysis potential.
• The creation of the spatial units is arbitrary,
  modifiable and subject to the whims and fancies
  of whoever is doing or did the aggregating
  (Oppenshaw, 1984).

31
MAUP

• Large regions could be divided into smaller
  spatial units in a large number of different
  ways.
• Aggregation of 1000 units to 20 results in 101260
• Aggregation will provide different outcomes.

32
MAUP

• Census collected by essentially non-modifiable
  units
• individuals
• households
• Reported by modifiable units.
• Basis of spatial units questionable
• historic
• workload
• political
• administrative

33
MAUP

• If the spatial unit basis is questionable, then
  any research work is potentially invalid.
• Does this matter?
• Yes
• The type and size of the areal unit can mask or
  enhance characteristics.

34
MAUP

• Scale Problem
• data for one area are progressively aggregated
  into fewer and larger units for analysis.
• Aggregation Problem
• alternative combinations of spatial units at
  similar scales
• Ecological Fallacy
• Closely related to MAUP
• results based on aggregate data can be applied to
  individuals within the zones.

35
MAUP

• A simple exercise

ASD Example
36
Gehlke and Biehl (1934)

• correlation coefficient increased with spatial
  unit aggregation.
• Eg. 252 units -0.502
• 25 units -0.763
• Concluded that a high correlation might occur by
  census tract when the traits were completely
  dissociated with the individuals or families

37
Yule and Kendall (1950)

• Agricultural Eg.
• Wheat and Potato yields correlation coefficients
  tended to increase with scale.
• No of Areal Units Correlation
• 48 .2189
• 3 .9902
• Concluded ..we seem able to produce any value
  of correlation from 0 to 1 merely by choosing an
  appropriate size of the unit to measure the yield

38
Solutions ?

• Ignore and hope the results are in some way
  meaningful.
• This result is implicit in the lack of statements
  accompanying most geographical work.

39
Solutions ?

• Accept normal science view that the zoning unit
  should be independent of the phenomena they are
  used to report.
• Identify meaningful spatial objects
• different units for different purposes

40
Solutions ?

• Design of unit not separate from analysis.
• Hypothesis
• identify spatial unit for the desired outcome
• what do results mean ? Support or dispute
  hypothesis - go back to step 1
• constraints imposed
• satisfactory outcomes?
• Go back

41
Solutions

• Ensure the problem is acknowledged in the text.
• Consider the impacts.
• Be a smart user of these data.
• Look at possible areal unit clusters, to better
  represent your outcomes.

42
Using Social Data

• Covered a range of issues..
• Data integrity.
• Spatial Accuracy.
• Spatial Change over time.
• MAUP
• Notwithstanding these issues, how are these data
  used?

43
Profiles

• What is a profile?
• Focus on subject matter.
• Focus on area(s).
• Example
• Womens Health 2000
• Onkaparinga Community Profile

44
Profile

• What would you include in a profile?
• Subject based?
• Area based?

45
Profile

• Population total and change
• age structure and change
• Ethnicity
• Income
• Housing tenure
• Mortgage and rental
• Family type
• Dwelling Structure
• Occupation
• Industry
• Education
• Car ownership

46
Atlas

• What would you include?
• Subject based?
• Area based?

47
Profile or Atlas

• These are often viewed as same thing.
• Eg.,
• Atlas of Australian People
• Social Atlas of Onkaparinga
• Country Matters Social Atlas of Rural and
  Regional Australia
• A Social Health Atlas of Australia

48
Atlas

• Whatever there name these reports are a simple
  but powerful means of providing social statistics
  and indicators.

49
Health Atlas

• The information in the atlas adds to a convincing
  body of evidence built up over a number of years
  in Australia on the striking disparities in
  health that exist between groups in the
  population. People of low socioeconomic status
  (those who are relatively socially or
  economically deprived) experience
  worse health than those of higher socioeconomic
  status for almost every major cause of mortality
  and morbidity. The challenge for policy makers,
  health practitioners and governments is to find
  ways to address these health inequities.

50
Health Atlas
51
(No Transcript)
52
(No Transcript)
53
Profile

• Onkaparinga Example

54
Onkaparinga

• FAMILIES
• One Parent Families with Dependent Children 
• Couples with Dependent Children 
• DINKs (double income, no kids) 
• INCOME
• Low Income Households
• High Income Households 
• LABOUR FORCE
• Unemployed People 
• Unemployed People Aged 15-24 Years 
• Unemployed People Aged 25-34 Years 
• Unemployed People Aged 45-64 Years 
• Managers, Administrators and Professionals 
• Mothers in the Labour Force 
• People who Travelled to Work by Car 
• People who Travelled to work by Public Transport 
• DWELLINGS
• Owner-Occupied Dwellings 
• Dwellings Being Purchased

• POPULATION
• People Aged 0-4 Years 
•  People Aged 5-14 Years 
•  People Aged 15-24 Years
• People Aged 60-74 Years
• People Aged 75 Years or More 
•  People Counted at a Different Address 
• ETHNICITY
• Aboriginal and Torres Strait Islander People 
• People Born Overseas 
• People Born in the United Kingdom or Ireland
• People Born in Southern Europe
• People Born in South East Asia 
• People not Fluent in English 
• Recent Arrivals 
• EDUCATION
• People with University Qualifications 
• People with Skilled Vocational (Trade)
  Qualifications 
• People without Qualifications 

55
Onkaparinga
56
Onkaparinga
57
Onkaparinga
58
Onkaparinga
59
Onkaparinga
60
Onkaparinga
61
Questions