Designing Zoning Systems for Research, Policy, and Data Reporting

Click here to start

Table of Contents

Designing Zoning Systems for Research, Policy, and Data Reporting

Designer Zoning Systems

The Results you obtain depend on the Definition of the Zoning Systems you use

The Results you obtain depend on the Definition of the Zoning Systems you use

You just need to be CAREFUL in how you handle Zone Design

Zoning Systems are very important and their design cannot be left to chance

Its a HARD problem because it involves the manipulation of both SCALE and AGGREGATION components of spatial data

andthere is little useful theory to guide the process

Spatial Patterns are SCALE and AGGREGATION sensitive!

How is it handled at present?

Badly is the short answer

So How Is Zone Design handled at present?

A major source of variation is not under control!

Map of Zones used by EU

Scale AND Aggregation are inseparably interlinked

Some Definitions

Some Observations

Spatial Analysis and Modelling assume you know the scale at which patterns or relationships exist

There are very few spatial analysis tools which are multi scaled and multi-aggregational

Important to start to DESIGN zones to MANAGE the scale and aggregation components directly!

Example using 1991 census data for Leeds - Bradford region

Unemployment fallacies for wards

A new approach!

Zone Design can also be used as a Spatial Analysis and Modelling tool

How do you DESIGN ZONES?

Handling the AGGREGATION component

ZDES Algorithms

Additionally Constrained ZDES

These design functions are best handled as inequality constraints so the zone design problem is analogous to a mathematical programming formulation

Unconstrained Zone Design

Unemployment Positively r=+0.99 correlated with Ethnic Minorities

Unemployment negatively correlated r=-0.999 with unemployment

Constrained Shape Equal Population Zoning

Unemployment Equal Economically Active Population

Unemployment Equal Population Zones with Shape Constraint

Policy Case Study: DOE’s Deprivation Score

Ward level indicators

What DAMAGE did the use of WARDS do to the results?

The 1991 Deprivation Index Signed Chi-squared function

Leeds - New Zoning

Leeds 3worst wards

Leeds 3 worst new regions

Some experiments with designing output areas for the 2001 UK census

Traditional Census EDs ?

So, what has changed ?

A good census output geography?

Designing Census Output Areas

Some experiments

An automated census output area design process

How do you OPTIMIZE for scale effects in a geographical analysis problem?

Multiple ZDES runs

BUTwhat do you do with all the output?

An example of looking for optimal scale patterns in spatial analysis of long term limiting illness data in Northern England

Ward Level map of Long Term Limiting Illness

GAM/K cluster identification

The hope is that some patterns or relationships or models will show chaotic behaviour as scale is changed

Affective Spatial Analysis?

It would be a ..

It would be a BIG mistake if we did not start to utilize the NEW opportunities

Author: Stan Openshaw, University of Leeds

Email: stan@geog.leeds.ac.uk

Home Page: http://www.geog.leeds.ac.uk/staff/s.openshaw/