Interpreting Maps

1

• Interpreting Maps

Dyfed Thomas Senior Public Health Information
Analyst NHS Kingston
2
Interpreting Maps
3
Outline

• Mapping Basics
• Colour Selection Data Classification
• Issues to Consider

4
Thematic Map Types
Dot map

• Proportional symbols

Charts as Symbol
1. Mapping Basics
5
What Does A Thematic Map Show?

• Value or magnitude of attribute data
• Spatial location and distribution

• Correlation amongst data if a bivariate /
  multivariate map

1. Mapping Basics
6
Scale, Colour Classification

• Scale is very important
• Consider how much land area is represented on
  your map?
• Number of colours graduation of colour
• What data classification is used?

2. Colour Selection and Data Classification
7
Selections

• Light-to-dark

Spectral schemes
2. Colour Selection and Data Classification
8
Classifications Available

• Quantiles (4,5)
• Equal Interval

• Equal Range
• Standard Deviation

2. Colour Selection and Data Classification
9
Ecological Fallacy MAUP

• Ecological Fallacy
• Uniformity
• Modifiable Areal Unit Problem (MAUP)
• Scale (or Size)
• Shape

3. Issues to Consider
10
EF - Uniformity

• Ecological fallacy derives from ascribing to
  individuals the characteristics of a group
• Consider the mean height of children in school
• 1.22, 1.14, 1.54, 1.67, 1.12, 1.28
• Mean 7.97/6 1.32
• No one in the class is actually this height
• So how representative is the mean height?

3. Issues to Consider
11
EF - Uniformity
Lower Super Output Area IMD
Ward Index of Multiple Deprivation (IMD)
3. Issues to Consider
12
What is MAUP?(Modifiable Areal Unit Problem)

• It is a potential source of error that can effect
  spatial studies which utilize aggregate data
  sources.
• Occurs when aggregation units are arbitrarily
  chosen.
• A completely new and different set of boundaries
  could display the base data in an equally
  meaningful way.

3. Issues to Consider
13
MAUP Effects of Scale (or Size)

• The two diagrams both contain the same data
  totals, and illustrate the change in mean score
  as smaller units are aggregated to form fewer
  larger units

9 LSOA
4 Wards
3
4
2
(3 4 5 6)
(2 7)
18
9
5
6
7
(1 8)
1
8
9
9
9
Mean Unit Value 5
Mean Unit Value 11.25
Source White P. (2005) Confronting the
Ecological Fallacy and the Modifiable Areal Unit
Problem A Geographical Information Systems
Approach to Improve Spatial Analysis and
Cartographic Visualization for Epidemiology
Unpublished PhD Thesis
3. Issues to Consider
14
MAUP Effects of Shape
Both areas contain the same total, the same
number of regions, and the same mean score, and
illustrate the change in data distribution as
boundaries of the regions are altered
3 Wards
3 Wards
(9 - 1)
(18 9)
8
28
9
9
18
18
(1 18 9)
Mean Unit Value 15
Mean Unit Value 15
Source White P. (2005) Confronting the
Ecological Fallacy and the Modifiable Areal Unit
Problem A Geographical Information Systems
Approach to Improve Spatial Analysis and
Cartographic Visualization for Epidemiology
Unpublished PhD Thesis
3. Issues to Consider
15
MAUP Example
Area outside of range of deprivation for inclusion
3. Issues to Consider
16
MAUP Example Legionella Outbreak
Census Area Unit Cluster Map
Meshblock Area Cluster Map
Source White, P., Janssen, K., Mason, K (2005)
Spatial Epidemiological Investigation of
Legionellosis in Christchurch. PHI
3. Issues to Consider
17
Summary

• Presenting and interpreting maps is an art as
  well as a science
• Be aware of purpose Are emotive colours being
  used to over-emphasize?
• Look at how data is classified
• All spatial analysis is subject to EF and MAUP
  you must keep it in mind
• Using maps is an important tool

18
Online atlases

• Cancer atlases www.ncin.org.uk/analysis/eatlas.s
  html
• Child Maternal Health Observatory
  www.atlas.chimat.org.uk
• Diabetes InfoBank (Diabetes UK) -
  http//www.diabetes.org.uk/InfoBankUK/
• Health Profiles www.apho.org.uk/default.aspx?RID
  49802
• Hospital Episode Statistics www.lho.org.uk/LHO_T
  opics/Analytic_Tools/InteractiveMaps.aspx
• Joint Strategic Needs Assessment -
  http//www.swpho.nhs.uk/resource/view.aspx?RID350
  85
• Obesity atlases www.noo.org.uk/maps
• Programme budgeting atlases - nww.nchod.nhs.uk,
  go to Compendium Indicators (not the latest data,
  but NCHOD are planning to update these)
• Teenage pregnancy www.empho.org.uk/datastore/map
  s/tpwardmaps.aspx www.empho.org.uk/THEMES/teenagep
  regnancy/tp11.aspx