GIS Lecture5

1

Attribute Data Models

GIS Database Management System

Attribute Data Models

GE517Engr. Ablao

Introduction

GIS involves both spatial and attribute data.Spatial – geometry of map featuresAttribute – characteristics of the map features

Attribute data are normally stored in tables.Record or tuple – rowField or item – columnA ib i i f d lAttribute – intersection of row and column

Data models relate spatial & attribute data.

8/20/2010GE 517 Geographic Information System

2

Spatial data (left) are linked to attribute data (right) by the label ID.


File Structures (File-based datasets)Simple list

Simplest file structureUnordered/unstructuredArrangement is by whichever comes first

Ordered sequential filesSimple lists that are arranged according to some order (ex. Alphabetical order)

Indexed filesAn index to the directory is needed for more efficient searches involving finding entries given certain criteriaCan be developed as direct files or inverted files


3

Indexed FilesDirect Indexed Files

Records are used to provide access to other ppertinent information

Indirect Indexed FilesIndex is based on possible search criteria, not on the entities themselvesAttributes are the primary search criteria and the entities rely on them for selectionentities rely on them for selection


Flat file databaseContains all data in a large file

Software could only operate on one file at a timey p

Format is very inflexible with respect to the modification of the database structure


4

Flat file database


DatabaseAn integrated set of data on a particular subject

Collection of interrelated data stored together with controlled redundancy to serve one or more applications in an optimal fashion

Requires more elaborate structure called a database structure or database management system

• A DBMS manage attribute data in separate tables


5

Significance of DatabaseMost GIS activities consist of storing entity and attribute data so that we can retrieve any combination of these objects.y j

Each graphical feature must be stored explicitly with its attributes so that their combined search becomes faster.


Advantages of Database over File-based datasets

Collecting data at a single location reduces redundancy and duplicationp

Lower maintenance cost due to better organization and decreased data duplication

Multiple applications can use the same data and can evolve separately over time


6

Advantages of Database over File-based datasets

User knowledge can be transferred between applications more easily because database remains constantFacilitated data sharing, with a corporate view provided to data managers and usersSecurity and standards for data and data access can be established and enforced


Types of Database Structure1. Hierarchical Data Structures

2. Network Systemsy

3. Relational Database Structures


7

Hierarchical Data Structure‘one-to-many’ or ‘parent-child’ relationship

Implies that each element has a direct relationship to a number of p psymbolic childrenEach child is capable of having the same direct relationship with his/her own offspring, and so on.


Hierarchical database


8

Hierarchical Data StructureAdvantages:

Simple and straightforward data access since parent and p g pchildren are directly linked

Easy to search since structure is well defined

Relatively easy to expand by adding new branches and formulating new decision rules


Hierarchical Data StructureDisadvantages:

Confined to queries along one branch onlyq g y

Difficult restructuring to allow other possible search criteria

Creates large index files

Redundant entries for searching


9

Network Systems‘many-to-many’ relationship

Each individual data is linked directly to anywhere h d b h hin the database using pointers, without the parent-

child relationship.


Network database


10

Network Systems


Network SystemsAdvantages:

Less rigid compared to hierarchical structureg p

Can handle many-to-many relationships

Allows much greater flexibility

Reduced redundancy of data


11

Network SystemsDisadvantages:

In very complex GIS, the number of pointers can b l th i i l t f t become large, thus requiring a lot of storage spaceLinkages between data must still be explicitly defined using pointersNumerous possible linkages can become extremely tangled, resulting to confusion and incorrect linkagesgNot recommended for novice users


Relational Database Management Systems (RDBMS)

Data are stored as ordered records or rows of attribute values called tuplesTuples are grouped with corresponding data rows in a form called relationsEach column represents data for a single attribute for the entire dataset


12

Relational Database Management Systems (RDBMS)

A key represents one or more attributes whose values can uniquely identify a record in a table.A k bl bli h A key common to two tables can establish connection between records in the tables.

Primary key – a column which is used to define the search strategy or criterion

Foreign key – column in the second table to which the primary key is linkedwhich the primary key is linked


Relational database


13

Relational Database Management Systems (RDBMS)Advantages:

Allow us to collect data in reasonably simple tables, keeping y p p gorganization also simpleCapable of doing relational joins, as long as there is at least one column common to the tables to be joined Allows greatest flexibility, both in design and querying


Normalization of relational databaseNormalization is a process of decomposition, taking a table with all the attribute data and breaking it down to small tables while breaking it down to small tables while maintaining the necessary linkages between them.Normalization is designed to avoid redundant data in tables, to ensure that attribute data in separate tables can be maintained and updated separately and can be linked when necessary, and to facilitate a distributed databaseto facilitate a distributed database.Normalization slows down data access.


14

PIN Owner Address Sale date Hectares Zone code Zoning

P101 Gloria 101Pampanga St.

01-20-2001 1.2 1 Residential

Erap 202San Juan St.

P102 Fidel 303Pangasinan St.

06-30-1992 1.5 2 Commercial

Cory 404Tarlac St.

P103 F di 505 06 30 1965 2 1 2 C i lP103 Ferdie 505Ilocos Norte St.

06-30-1965 2.1 2 Commercial

P104 Dado 606Pampanga St.

06-30-1961 0.8 1 Residential

Unnormalized table


PIN Owner Address Sale date Hectares Zone code Zoning

P101 Gloria 101Pampanga St.

01-20-2001 1.2 1 Residential

P101 Erap 202San Juan St.

01-20-2001 1.2 1 Residential

P102 Fidel 303Pangasinan St.

06-30-1992 1.5 2 Commercial

P102 Cory 404Tarlac St.

06-30-1992 1.5 2 Commercial

P103 F di 505 06 30 1965 2 1 2 C i lP103 Ferdie 505Ilocos Norte St.

06-30-1965 2.1 2 Commercial

P104 Dado 606Pampanga St.

06-30-1961 0.8 1 Residential

First Normal Form


15

Second NormalForm

8/20/2010 GE 517 Geographic Information System

NormalizedNormalizedForm

8/20/2010 GE 517 Geographic Information System

16

Data Storage in a DBMSObject classes/layers are stored in database tablesEach layer is stored as a single database table in a database management systemRows contain objects, while columns contain attributes/properties of the objects


Basic Database Functions/OperationsJoin

Tables are joined together using common row/column values or keysj g g yAfter joining two or more tables, a new table is created which contains all the values of the joined tables

Database tables can be joined together to create new relations, or views of the database.


17

Basic Database Functions/OperationsLink

Tables are linked using common row/column values or keysg yUnlike in joining, linking tables does not result to a new table. The original tables are retained but accessing one enables the user to also access a table linked to it


Database DesignInvolves three stages: conceptual, logical, and physical

Involves six practical steps (see Figure)p p ( g )


18

Stages of Database DesignConceptual Model

Logical ModelUser View

Object and

Relationships

Logical Model

Geographic Database

Types

Geographic

Physical Model

Database Schemap

Geographic Representation

Geographic Database Structure


Conceptual ModelSteps involved are:

1. Model the user’s viewIdentifying organizational functions, determining data requirements of these functions, organizing data into groups for data managementMay be presented using a report with tables


19

Conceptual Model2. Define objects and their relationships

Specification of object types/classes and functions, and their p yprelationships May be presented using diagrams



20

Conceptual Model3. Select geographic representation

Choosing between the types of discrete objects (point, line, or polygon) or field to represent the (point, line, or polygon) or field to represent the dataSelection has a critical impact on the database useAlthough it is possible to switch between representations later on, it would be computationally expensive and would lead to information loss


Logical ModelSteps involved are:

1. Match to geographic database typesMatching of object types to be studied to specific data types supported by the GIS

2. Organize geographic database structureDefining topological associations, specifying rules and relationships, and assigning coordinate systems


21

Physical ModelStep involved is:

Define database schemadefinition of the actual physical database schema that will hold the database data valuesusually created using the DBMS software’s data definition language (ex. SQL)


Attribute data entryField definition

Attribute data entryy

Attribute data verification

Creation of new attribute data


22

Field definitionDefinition of (a) field name, (b) data type, (c) data width, and (d) number of decimal places.

Data type may be (a) numeric (integer or floating-point), (b) string, (c) Boolean, or (d) date.

Consider measurement scale of data.


Attribute data entryAkin to digitizing for spatial data entry

Attribute data need to be entered by typingGiven: map with 2,000 polygons and 10 fieldsTime: At 10 seconds per value, it takes 55 hours – 33 minutes – 20 seconds (2.3 days) to enter 20,000 values

Best to determine if an organization has attribute data in digital format (e g NSO)attribute data in digital format (e.g. NSO)


23

Attribute data verificationIn this step:

Ensure attribute data are properly linked to spatial dataEnsure attribute data are properly linked to spatial dataVerify the accuracy of attribute data

May be difficult due to observation errors, out-of-date data, and data entry errors

To check for errors:Table may be printed for manual verificationy pComputer programs may be written to automate task


Creation of new attribute dataAttribute data classification

Example: ElevationHigh = {Higher than 600 meters}

Medium = {Between 200 and 600 meters}

Low = {Lower than 200 meters”

Attribute data computationExample: Soil erosion potential = rainfall parameter × Soil parameter ×topographic parameter × land cover parameter × management parameter

Example: Agricultural harvest = area × potential yield


GIS Lecture5

Documents