IntroductionTools

Applications

New tools to create a W matrix that representsthe socioeconomic reality of the study data.

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2)

(1) Rise-group (Department of Economics - EAFIT University)(2) ESRI

ERSAAugust, 2011

Barcelona

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

Outline

1 Introduction

2 ToolsClusterPyGeoGrouperArcGIS

3 ApplicationsApplication 1: Constructing W for spatial modelApplication 2: Getting rid of spurious spatialautocorrelation

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

Outline

1 Introduction

2 ToolsClusterPyGeoGrouperArcGIS

3 ApplicationsApplication 1: Constructing W for spatial modelApplication 2: Getting rid of spurious spatialautocorrelation

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

Outline

1 Introduction

2 ToolsClusterPyGeoGrouperArcGIS

3 ApplicationsApplication 1: Constructing W for spatial modelApplication 2: Getting rid of spurious spatialautocorrelation

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

W matrices: types of representation (Based onAldstadt and Getis, 2006)

Type I: Theoretical notion of spatial associationThere exist some theory behind its definitionIt is exogenous to the model

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

W matrices: types of representation (Based onAldstadt and Getis, 2006)

Type II: Geometric indicator of spatial nearnessExogenousUseful when absence of theoryThe only assumption here is that near areas are morerelated than areas that are further apart.

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

W matrices: types of representation (Based onAldstadt and Getis, 2006)

Type III: Descriptive expression of the spatial association withina set of data

Modeler allows study data to speak for themselves.Extract the spatial relationships from the data.Ensure that the complexity of spatial association withintheir data will be included in the model.Types I and II: explanatory models; and type III: descriptivemodels.

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

W matrices: types of representation (Based onAldstadt and Getis, 2006)

Type III: Descriptive expression of the spatial association withina set of data

There exist many algorithms in the literature that can beutilized to construct this type of matrices.Most of them are just in the papers, its implementation isnot available to other researchers.

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

Objectives

At leastTo pull the methods out of the papers into the computer

Ultimate goalTo make a development one standard application to apply,customize and compare algorithms in a manner that has notexisted before

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

Tools for constructing type III W matrices

DescriptionAlgorithms for spatially constrained clustering.

Identify spatial association among nearby units.Weights matrix that is representative of the observedspatial association in the data.

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

ClusterPyGeoGrouperArcGIS

Outline

1 Introduction

2 ToolsClusterPyGeoGrouperArcGIS

3 ApplicationsApplication 1: Constructing W for spatial modelApplication 2: Getting rid of spurious spatialautocorrelation

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

ClusterPyGeoGrouperArcGIS

ClusterPy

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

ClusterPyGeoGrouperArcGIS

ClusterPy

DescriptionIt was released in June 2011Open sourceMultiplatformClusterPy is now included in the official repository ofsoftware for Python

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

ClusterPyGeoGrouperArcGIS

Available algorithms

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

ClusterPyGeoGrouperArcGIS

ClusterP

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

ClusterPyGeoGrouperArcGIS

ClusterPy

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

ClusterPyGeoGrouperArcGIS

ClusterPy

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

ClusterPyGeoGrouperArcGIS

ClusterPy

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

ClusterPyGeoGrouperArcGIS

ClusterPy

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

ClusterPyGeoGrouperArcGIS

ClusterPy

Forthcoming features

MPS file for max-p-regions problem.MPS file for p-regions problem (Duque, Church,Middleton).Module for experiments.Module for simulating irregular lattices (fractal theory +Comp. geometry + Stochastic calculus)Models for specific applications (Electoral districting,School districting, Turfing, Sales districting, health-caredistricting...)

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

ClusterPyGeoGrouperArcGIS

ClusterPy

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

ClusterPyGeoGrouperArcGIS

Outline

1 Introduction

2 ToolsClusterPyGeoGrouperArcGIS

3 ApplicationsApplication 1: Constructing W for spatial modelApplication 2: Getting rid of spurious spatialautocorrelation

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

ClusterPyGeoGrouperArcGIS

GeoGrouper

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

ClusterPyGeoGrouperArcGIS

GeoGrouper

DescriptionGUI version for ClusterPyFreewarePoint and clickMultiplatformTo be released in 2012

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

ClusterPyGeoGrouperArcGIS

GeoGrouper

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

ClusterPyGeoGrouperArcGIS

GeoGrouper

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

ClusterPyGeoGrouperArcGIS

GeoGrouper

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

ClusterPyGeoGrouperArcGIS

GeoGrouper

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

ClusterPyGeoGrouperArcGIS

GeoGrouper

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

ClusterPyGeoGrouperArcGIS

GeoGrouper

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

ClusterPyGeoGrouperArcGIS

Outline

1 Introduction

2 ToolsClusterPyGeoGrouperArcGIS

3 ApplicationsApplication 1: Constructing W for spatial modelApplication 2: Getting rid of spurious spatialautocorrelation

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

ClusterPyGeoGrouperArcGIS

ArcGIS Toolbox

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

ClusterPyGeoGrouperArcGIS

ArcGIS

DescriptionCommercialRuns on WindowsWill be available in the next release of ArcGIS

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

ClusterPyGeoGrouperArcGIS

ArcGIS Toolbox

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

ClusterPyGeoGrouperArcGIS

ArcGIS Toolbox

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

ClusterPyGeoGrouperArcGIS

ArcGIS Toolbox

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

Application 1: Constructing W for spatial modelApplication 2: Getting rid of spurious spatial autocorrelation

Outline

1 Introduction

2 ToolsClusterPyGeoGrouperArcGIS

3 ApplicationsApplication 1: Constructing W for spatial modelApplication 2: Getting rid of spurious spatialautocorrelation

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

Application 1: Constructing W for spatial modelApplication 2: Getting rid of spurious spatial autocorrelation

Creating W with AMOEBA algorithm

Aldstadt J, Getis A (2006) Using AMOEBA to create a spatial weights matrixand identify spatial clusters. Geogr Analysis 38(4):327–343

Total fertility levels in Amman, Jordan.Spatial error formulation model using anAMOEBA-generated spatial weights matrix.

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

Application 1: Constructing W for spatial modelApplication 2: Getting rid of spurious spatial autocorrelation

Outline

1 Introduction

2 ToolsClusterPyGeoGrouperArcGIS

3 ApplicationsApplication 1: Constructing W for spatial modelApplication 2: Getting rid of spurious spatialautocorrelation

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

Application 1: Constructing W for spatial modelApplication 2: Getting rid of spurious spatial autocorrelation

Using Max-p regions model to control for spuriousspatial autocorrelation

Weeks JR; Hill A; Douglas S; Getis A and Debbie F (2007). Can we spot aneighborhood from the air? Defining neighborhood structure in Accra,Ghana. GeoJournal. 69(1–2): 9–22.

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

Application 1: Constructing W for spatial modelApplication 2: Getting rid of spurious spatial autocorrelation

Where to get these tools?

http://code.google.com/p/clusterpy/http://pypi.python.org/pypi/clusterPy/0.9.9www.rise-group.org

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix

IntroductionTools

Applications

Application 1: Constructing W for spatial modelApplication 2: Getting rid of spurious spatial autocorrelation

Thanks.

Duque JC (1), Janikas MV (2), Rosenshein L (2), Scott L (2) Tools to create a W matrix