Longley et al., ch. 8

Geographical Data Modeling

UML and Data Modeling ElementsExamples from the Marine Data Model and ArcHydro

(Thanks to Dawn Wright)

Longley et al., ch. 8

Models for GISrepresentation of reality --> model

GIS itself is based on a model of complexity and used to model complexity

Not a full representation of reality: even at 1:1

data model = limited representation of reality a discretization or partitioning of space finite, discrete nature of computing devices

Data Model

Representation of information about a form or a process

E.g. a weather map with isotherms

FIELD … “grid”

OBJECT … “isoline”

A good model allows us to infer process

FIELD or OBJECT … “geo-relational”

3 Data Models

geo-relational coverage (object view; from classic ArcInfo polys)

geo-relational shapefile (object view; from ArcView)

geodatabase (object-oriented; new in ArcGIS 8,9 etc.)

Object Oriented GIS (p. 191)

Object: a self-contained package of information describing an entity

Collection of objects – class

Objects can have behavior – encapsulation

Inheritance – reusable objects

Polymorphism-objects can have their own implementation for application-e.g. create, draw, delete

A

B

command line interfaceUnix, NT, WindowsArcView as interface

Georelational Data Model:“Classic” ArcInfo and ArcView

ArcInfo Coverage as a Data Structure

Image courtesy of Louisville/Jefferson County Information Consortium, Kentucky

Data StructureNot tied to process at allConcerned simply with what can be computed and what can’tWay in which the data model is represented in the GISDEM for grid or layer model (e.g. array)contour for isoline modelTIN for Delaunay triangulation modelcoverages, shapefiles for geo-relationalgeodatabase for OO geodatabase

ArcGIS 9 Iconsyellow coverage

gray geodatabase

green shapefile

Geodatabase

Features and attributes as objectsRelationships among features encodedValidation or editing rules, behaviors“Container” for Vector, raster, tabular

data Relationships Topology

MS-Office

Relationships for a feature

ArcGIS Data Modelssupport.esri.com/datamodels.cfm

ArcMarine dusk.geo.orst.edu/djl/arcgis

Data Modeling for Spatial Analysis

What is spatial analysis? "a set of methods whose results change when the

locations of the objects being analyzed change"

Methods for working with spatial data to detect patterns, anomalies to find answers to questions to test or confirm theories

deductive reasoning-general to specific to generate new theories and generalizations

inductive reasoning-specific to general

What is Spatial Analysis (cont.)

Methods for adding value to data in doing scientific research in trying to convince others

A collaboration between human and machineHow do we set up the framework for spatial analysis? Data model to data structure

A Georelational to a Geodatabase Model

Coverage and shapefile data structures homogenous collections of points, lines, and

polygons with generic, 1- and 2-dimensional "behavior" as operations

Can’t distinguish behaviors Point for a marker buoy, same as point for

observation“smart features” in a geodatabase lighthouse must be on land, marine mammal siting must be in ocean

Objects can “self-police”

Basic template for implementing GIS projects input, formatting, geoprocessing, creating

maps, performing analyses

Basic framework for writing program code and maintaining applicationsdevelopment of tools for the community

Promote networking and data sharing through established standardscommon model=interoperability

Purpose of ArcHydro, ArcMarine etc.

ArcMarine Design Strategy

“Generic”

Marine Data Model

User Group

Data Model

User Group

Data Model

User Group

Data Model

Project

Data Model

Project

Data Model

Project

Data Model

Inh

erit

ance

Geodatabase ConceptsESRI's data object-oriented data model objects, features, behaviors

Geodatabase collection of feature data sets, rasters, TINs all data in relational tables behavior is coupled with features through rules

(object-orientation) Supports model-builder for processes

Feature data set contains feature classes defines topological role of features has a coordinate system

Geodatabase Concepts ( cont. )

Feature classstored in a relational tablespecial field for geometric shape geometric data incorporated into the

databasePoint, multipoint, segment, path, ring, polyline,

polygon

Geodatabase Feature Class Geometries

Real World Objects and relationships

DatabaseSchema

(Object state)

Physical Model

Modeling ProcessConceptual Model

Lists, flow diagrams, etc

Logical ModelDiagram in CASE

Tool

Graphic courtesy of ESRI

Data Model Levels

IncreasingAbstraction

RealityReality

Conceptual ModelConceptual Model

Logical ModelLogical Model

Physical ModelPhysical Model

Human-oriented

Computer-oriented

Specific Steps in Data Modeling

(1) Conceptualize the user's view of data what are the basic features needed to solve the

problem?

(2) Select the geographic representation points, lines, areas, rasters, TINs

(3) Define objects, features, and relationships draw a UML diagram, specify relationships,

“behaviors”

(4) Match to geodatabase elements Refine relationships, “behaviors”

(5) Organize geodatabase structure, add data

( 1 ) User’s View of Data

( 1 ) User’s View of Data cont.

(2)Select

geographic rep.

Steps in Data Modeling(1) Conceptualize the user's view of data

what are the basic features needed to solve the problem?

(2) Select the geographic representation points, lines, areas, rasters, TINs

(3) Define objects and relationships draw a UML diagram, specify relationships,

“behaviors”

(4) Match to geodatabase elements Refine relationships, “behaviors”

(5) Organize geodatabase structure, add data

Unified Modeling LanguageEntity-relationship diagramsDesign the methodologies, diagram notations

UMLNot a design methodologyJust a diagrammatic notation based on

methodsEndorsed by leading software and database

companies

UML ( cont. )

Diagrammatic notation = “visual language”...

For constructing a data model

Drawings, relationships constructed in Visio (other tools available)

Tools to input a drawing into ArcGIS input drawing to the data model

UML Notation

a class is shown as a box

top part contains the name of the class

lower part contains the attributes

methods associated with the class

lines connect boxes and indicate relationships

Graphic courtesy of Maidment et al., ArcHydro team

UML Notation ( cont. )

Abstract class specify subclasses

underneath Mammals w/human or

dog feature classes no new instances

Feature Class Specify subtypes

underneath Human, dog, cat

Objects and Features

Object (real world) in ArcGIS an object is non-spatial it is NOT a point, line, or area it has no geographic location it has no shape attribute in its table Drainage network, ship, vehicle, … customer,

lake, house, etc.

Feature (spatial context) an object that has geographic location a point, line, area, TIN, raster

Relationships

Links between classes, shown as lines

One to one

One to many

Many to many

Relationships (cont.)

1:1 - solid line one record in Class A linked to one record in Class

B “is married to” the class of state capitals linked to the class of states

1:n - solid line with * at one end one record in Class A linked to any number of

records in Class B "owns" the class of states linked to the class of area codes

Relationships (cont.)

m:n - solid line with * at both endsany number of records in Class A linked to

any number of records in Class B"has visited”"was never married to" the class of mountain lions linked to the class of

wilderness areas

Graphic courtesy of Maidment et al., ArcHydro team

Type Inheritance

White triangleClass B inherits the properties (attributes, methods) of Class A

the class street inherits from the class transportation network

Solid diamondthe parts and the whole depend on each other

Graphic courtesy of Maidment et al., ArcHydro team

Steps in Data Modeling(1) Conceptualize the user's view of data

what are the basic features needed to solve the problem?

(2) Select the geographic representation points, lines, areas, rasters, TINs

(3) Define objects and relationships draw a UML diagram, specify relationships,

“behaviors”

(4) Match to geodatabase elements Refine relationships, “behaviors”

(5) Organize geodatabase structure, add data

Steps in Data Modeling(1) Conceptualize the user's view of data

what are the basic features needed to solve the problem?

(2) Select the geographic representation points, lines, areas, rasters, TINs

(3) Define objects and relationships draw a UML diagram, specify relationships,

“behaviors”

(4) Match to geodatabase elements Refine relationships, “behaviors”

(5) Organize geodatabase structure, add data e.g., Marine Data Model tutorial