GIS DATABASES

an overview
 GIS Databases - some issues (I)

• GIS data relates to geographic features or objects such as wells,

districts or roads. This data consists of the position of the feature
(spatial data) and its other attributes. Thus GIS data can either be
spatial (graphic/map) or attribute (textual). Currently, in most GIS
software packages the two data types are stored in separate
databases. The spatial data is usually stored and managed by a
proprietary software while the attribute data is stored in some
commonly available database package such as dBASE, ORACLE or
INGRES.

Screen 1 of 3

2
 GIS Databases - some issues (ii)

• The creation of a database represents an attempt at a structured and

systematic effort to collect useful data which can be accessed in
different ways to help users make decisions. A data dictionary is an
important tool which can be used to achieve this objective by
controlling and documenting data globally. A data dictionary is the
critical component of a complex piece of software commonly termed
a DBMS (database management system). (A related but wider
concept is a metadata).

Screen 2 of 3

3
 GIS Databases - some issues (iii)

• Relational databases are the most common type of database

structures in use today although some applications are based on
hierarchical structures. Relational databases are popular largely
because of their flexibility in design and their ability to support ad hoc
queries. In a relational database data is stored in tables which can be
linked if they have fields (keys) in common – considerable attention
has to be paid to these linkages in the design phase. This flexibility
can produce inefficient arrangement of tables or a data structure
which can produce serious data management problems over time.
Normalisation is one of several techniques for deriving an efficient
storage arrangement – it helps in identifying linkages and reduces
data duplication.
Screen 3 of 3

4
 Contents

– the basics of data storage

– overview of databases
• the database approach
• types of databases
– GIS database design considerations
– development of an ARC/INFO database
– GIS over the Web ( later in course)

5
 A storage hierarchy ...

– files/tables
• records
• fields(types …)
increasing
– databases complexit
y
– information systems
– decision support systems (DSS/SDSS)

– approaches to storage
• manual, application/file based, databases
Application based approach (urban
example)
Tax/Rates
Assessment Assessment Data

Permits Permit Data

Sewer
Sewer Data
Maintenance

Applications using data stored as Application Specific data

[Duplication, inconsistency, lack of sharing, map base …]
Database approach (urban example)
Streets

Database Management System

Addresses

Tax/Rates
Assessment Assessment Data

Permits Permit Data

Sewer
Maintenance Sewer Data

Database approach and use of shared data - Parcels

implications for GIS
 Database … a definition

A collection of interrelated data stored

together with controlled redundancy to serve
one or more applications in an optimal
fashion.
A common and controlled approach is used
in adding new data and modifying and
retrieving existing data within the data base
(Martin 1978)

10
Databases… objectives/advantages
– “centralised” approach to data storage and
management … global view of data … data
dictionary
• standardisation of all aspects of data management
• integrity constraints … validation enforced
• reduced duplication
• multiple access / retrieval flexibility
• ...
– data base management system (DBMS)

11
 Database/s… data dictionary

– the most critical (?) element of a database

– essential for system development … data
about data … metadata
– uses include
• design - entities and data relationships
• data capture - entry/validation
• operations - program documentation
• maintenance (impact assessment of proposed
changes , est. of effort, cost …)

12
 GIS Metadata

(Robinson et. al. 1996)

Spatial Data Transfer Standard (SDTS) - Data Quality Report

15
(Chrisman 1997, p 78)
Database Management System (DBMS) …
key software modules

– a data description/definition module

• defines/creates/restructures
• enforces rules
– [a data entry-update module]
– a query module
• retrieval for queries, ad-hoc queries, simple reports
– a report writing program
– a high level language interface
– ...

16
 Stages of Database Development
– information systems plan for organisation
– system specification
– conceptual design
– physical design of the database
– database implementation
– monitoring/audit

[See Chrisman, Chap 10, esp Fig 10.1]

17
Organisational strategy and IT
NSW Land Information System (LIS)
– Problems/issues:
• rationalisation of land related information in government
agencies ...the removal/reduction of duplication
• introduction of economies in data capture, maintenance and
storage
• better (and wider) access to data
– Solutions:
• better data distribution mechanism (data format and location
transparent to user)
• uniform query language (SQL), coding and data interchange
standardisation ( … SDTS)
• reduction of data duplication
18
Chrisman
(1997,p 242)
 Database types -
a history
Hierarchical
databases
Evolution of Database
technology

Relational
databases
 Database types - hierarchical (i)
– lends itself to GIS use as data are often
hierarchical in structure e.g. municipality x
province x country
– records divided into logically related fields …
connected in a tree-like arrangement
– master field in each group of records …
pointers … updates require pointers to be
modified
– fast preset queries … ad hoc queries difficult or
impossible
21
Hierarchical Database
for the storage of Database types
county boundaries - hierarchical (ii)
in USA
COUNTRY (USA)

States

Counties

Boundaries

Nodes
Hierarchical Structure for a Cadastral
database
Hierarchical Structure for vegetation mapping

24
 Database types - relational (i)

– model developed from mathematics

– records and fields in a 2-dimensional table
– any field can be used to link one table to
another
– normalisation … redundancy/stable structure
– ad hoc queries SQL… modifications easy
– not very efficient for GIS …SQL3

28
Database types
- relational (ii)
Database types - (geo) relational (iii)

30
Hierarchical structure
Network structure

Relational structure
(part…)

31
 GIS Database Design Basics

• Some things to consider …

– Fundamentally different ways of conceptualising the
real world (discrete features/continuous fields)
– different DATA MODELS with different data structures
possible (NB: software implementation)
– Graphic (spatial) and textual(attribute) data to be
managed together
– Considerations specific to spatial/geographic data

32
 Database Design Considerations
specific to spatial/geographic data

• Large volumes of data

– data filtering and or compression techniques
• Geographic scale and resolution, extent, projection issues
e.g. scale affects map content
• Handling databases of different scales: join problems (e.g.
urban rural), merge problems (different themes)
• Data organisation: partitioning/tiling, spatial indexes
• Symbolisation, annotation and display problems
• Metadata for a complex data set

33
 Approaches to GIS Database
software design
– Early: develop a proprietary system
– Current: develop a hybrid system: proprietary
graphics + commercial DBMS for attribute data
(e.g. ARC/INFO)
– Future: develop a spatial DBMS from scratch

36
 Approaches to GIS Database
software design

Early Current Future

Systems Systems Systems
37
GIS databases … some problems (i)
– centralised risk
• centralisation demands better quality control
otherwise higher potential for disaster
– cost
• large DBMSs are expensive to design, implement and
operate
• piecemeal design is difficult
– complexity
• need to keep track of complex hardware and software
• need to keep track of graphical as well as attribute
data and the links

38
 GIS database problems - an
illustration

Cascading effects of change in a GIS database (ESRI 1989)

39
GIS database design: an example

41
 Objectives of design

– a good design results in a database which:

• contains necessary data (controlled redundancy)
• organises data so that different users access the
same data
• accommodates different views of the data
• distinguishes applications which maintain data from
those that use it
• appropriately represents, codes and organises
geographic features

42
 Design methodology (for ARC/INFO)

1. Model the users’ view

2. Define entities and their relationships

3. Identify representation of entities

4. Match to ARC/INFO data model
5. Organise into geographic data sets

….. Physical implementation

43
 1. Model the users’ view

– create a model of work performed by users for

which ‘location’ is a factor
• identify organisational functions
• identify the data which supports the functions
– organise data into sets of geographic features
• data function matrix
– high level classification of data
– interdependence of data and function
– difference between users and creators of data
44
 Example: Land development management
function
Land
Development
Management

Assessor Parcel is the

geographic unit output
from this process.
Data function
matrix …
an example
 2. Define entities and their
relationships

– entities: distinguishable objects which have a

common set of properties
• identify and describe entities
• identify and describe the relationship among these
entities
• document the process (diagrams, data dictionary)
• Normalise the data

47
Entity/relationship definition

48
Diagramming … entities

49
 Normalisation …
efficient storage structure and linkages

ENTITIES: Parcel Zoning Owner Ownership

Normalisation is the process of identifying a more efficient storage (table)

structure which organises data in terms of the underlying entities.

First Normal Form (1NF) ,Second Normal Form (2NF)

Third Normal Form (3NF)
 First normal form (1NF)

APN - Assessor Parcel Number 51

Second normal form (2NF)

52
Third normal form (3NF)
3. Identify representation of entities

– determine the most effective spatial

representation for geographic features
– consider whether:
• a feature might be represented on a map
• the feature will have different representations and
different map scales
• textual attributes of the feature will be displayed on
map products
• ...

54
 4. Match to ARC/INFO data model
– determine the appropriate ARC/INFO
representation for entities
• points, lines, polygons
– ensure complex feature classes are supported
• route comprised of sections which in turn are based
on arcs
• a region is composed of polygons
• event is a point or a line which occurs along a route
– others (e.g. GRID, TIN)

55
5. Organise into geographic data sets

– a more “technical” procedure dictated by

ARC/INFO requirements… to identify and
name the geographic data sets that will contain
the various entities:
• define the contents of geographic data sets
(coverages/themes, grids etc)
• name workspaces, geographic data sets, entities
and attributes
• add cartographic text and lookup tables

57