319

spatial data has...

Location

Multidimensionality

Scale dependency

to be sampled

Spatial / temporal autocorrelation

Uncertainty

Special analysis methods

Great volume

Many origins

high expense and maintenance

Relational Database Management

A table ina geodatabase for keeping attribute data that relate to spatial features.

Can be associated with behaviour.

Custom features,

multiuser editing,

enhanced topoloy and validation

attribute validation (domains, subypes)

scalable storage solutions

relationships

Unified Modelling Language

Database Management Systems

1 Data definition Language (DDL)

2 Data manipulation and query lanuage

3 Data model that supports: standard types, specialized data types

4 Must support multiple views of data

5 Security

6 Concurrency control

7 Database administration tools (DBA)

DB design tools

Programmable API

Data is the foundation of a GIS installatino

It determines:

What you can analyse

Where your analysis focuses

Types of analysis

Quality of your analysis

It includes a locational dimension to its description of real world things.

Spatial data can be:

multidimensional (link time, place, attributes)

voluminous

expensive and time consuming to update

compiled from multiple sources

requrie special analysis methods

Scale dependent

Sampled data

spatial autocorrelation

temporal autocorrelation

uncertain

Error: the unkonwn uncertainty due to systematic & human lmiitations

Randomness: some things can never be modelled directly like tiny variations in the surface

Vagueness: uncertianty associated with a spatial or attribute concept.

Map projection, datum, parameters...

Improper representation of objects

Primary measurements eg Surveying methods

Secondary data acquisition errors: instrument related, media related, human operator. like a scanner or something.

a vector data storage format for location, shape, attributes of features

Feature = spatial entity (vector, raster,TIN)

Object = nonspatial entity (ie table)

Feature Class = collectiopn of single geometric type of features.

A feature dataset is a collection of related feature classes that share a common coordinate system. Feature datasets are used to spatially or thematically integrate related feature classes

rules that give allowable values for a field type.

Can be coded values eg Class 1 Class 2 Class3 or a range of values 0-500m

Validation rules verify that the data a user enters in a record meets the standards you specify before the user can save the record

Maintain integrity of feaures and attribues

Are subtypes of feature/object attributes.

simple behaviour,

An association between 2 object classes or 2 feature classes.

Object class - define th cardinality between tables eg M:1, 1:M

Feature Class - eg; a pole can have 0, 1, or 2 transformers.other values are not permitted.

A table in a geodatabase to keep descriptive data that relate to spatial features.

Can be associated with a behaviour

eg: Object class of landowners is joined with a polygon feature class of parcels.

A table that stores relationships between 2 features in feature classes, or two objects in 2 tables.

eg: how a feature changes when its related object changes

Arc Spatial Database Engine

file:

single user

project based

1TB limit

Separate dataset files

ArcSDE:

Scalable client/server architecture

Multiuser access and editing

No limit to size

Data centralised to help admin, user access, backup etc..

Multiple editing

Enhanced topology and validation

Attribute validation (domains, subtypes)

Scalable storage solutions

Relationships

Custom featues

Planar enforcement concept.

eg various types of soil polygons cannot overlap

Arcnode & polygon arc topology

1. Arcs start with a Fromnode and end with a Tonode

2. Arcs have a direction (based on from/to nodes

3. Connecting arcs share a common node.

these rules allow paths to be determined through a netowrk of connected arcs. Can ensure polygon features are closd.

1. Polygons stored as a list of arcs (polylines) that make the boundary.

2. Polygons have a single label point that links to an attribute table.

3. Based on Ac direction, left/right polygons are identified

32.

How many topology objects are created

The rules used in each topology

Feature classes the rules are applied to

Cluster tolerances

Rankings of feature classes

Permissible exceptions to topology rules.

Simple edge - connecito to 2 junctions a ach end

Complex edge- connection to 2 junctions at each end wtih more juncitons in beween

User defined junction - defined from users feature data source

Orphan junction - created when edges are created.

1: Every node set a tentative distance value. Initial node = 0, every other node = infinit

2: Keeps two sets of nodes: visited & unvisited

3: For the current node, the distance to each neighbouring node is calculated plus the distance to the current node.

4: If this distance value is les than their tentative value (which begins at infinite), then replace.

5: When all neighbours checked, mark current node as visited, and remove from unvisited.

6: Set the unvisited node with the smallest tentative distance as the next current node.

7: Repeat from 3.

Finished if the destination node has been marked visited.

The location of events, want to know about the distribution of these points
(events)

Frequency / density

geometric centre

Spatial dispersion

Spatial arrangement - autocorrelation, clustering

Geometric centre and dispersion.

Geometric centre - mean of X and Y. dispersion = standard deviation.

A mean with large standard deviation is not a reliable indicator of the distribution o centre

Clustered

Scattered

Random

measures dispersion based on minimum interfeature distance.

The average distance between points in a clustered pattern is most likely less than in a scattered pattern.

measures the extent to which the occurence of one feature is influenced by the distribution of similar features nearby.

Occurs if the existence of one feature attracts similar features to its neighbourhood.

Hierarchical: start with all patters as single cluster. Splitting and merging until threshold met. results in ree or dendogram.

Partitional: eg kmeans or kmedoid result in spherical cluster.

Gridbased and density based types

object is a structure that represents a single entity.

Describes its information content (attribute/properties = its state) and its behaviour.

every object belongs to a class.

defines a structure and set of operations that are common to a group of objects.

an instance

State + Behaviour

Unifid Modelling Language

UML is a graphical modelling language

Supports object oriented design

Standard, widely used

Graphical characeristics

Geographic location

all other associated data

Encapsulation:

ability of an object to hide its internal structure

ensures data independence

Inheritance:

Object classes defined hierarchically

Generall classes define the structure of generic objects, then specialised by creating subclasses.

Subclasses get all parent properties/methods and may add their own.

NULL marker for missing info. this is best to avoid.

why? Because things then cant be meaningfully compared, theres issues wiht logical comparisons, why have an attribute if it can be missing???

A colleciton of zero or more associated attributes

A collecitn of structurally identical tuples

Tuple

Relation

Attributes

Look like tables

Tuples ina relation are unique

Attributes are single valued

Relations are not ordered

the number of attributes

the number of tuples

Only one per relation

Comprise of one or more attributes (composite vs not)

No nulls - no missing vlaues

No unnecessary attributes (if you have IRD, not much point having IRD + name)

Stable - not change over time

May be generated (surrogate)

Property - valuation number

Vehicel, registration

Student, student ID

zero or more per relation

comprise one or more attributes

must reference a OK in another relation, or must be completely missing

May or may not be unique.

Often change over time. (as papers change as customers change)

Can overlap with other FKs and PKs

Employee: department ocde, manager ID

Customer: salesrep ID

Lecture: paper code, room code

Ensure consistanc of data.

data values must: make sense, satisfy all specified business rules, avoid unnecessary double ups

Projeciton

REstriction

Join

Set union: intersection, difference

extracts some subset of the attributes of a relation.

duplicate tuples are automatically deleted.

extract some specified subset of the typles of a relation

combins associated tuples fom different relations

combine tuples from two instances of the same relation

tuples that appear in either one relation or the other

tuples that appear in both relations at once

tuples that appear in one relation bt not the other

Is predominant abstract model for attribute data.

Robust, flexible, theoretically sound.

Strong data integrity protection.

Basis for all modern database management systems.

Database Management System is a computer system that allows users to access a database without knowledge of how data are studied physically.

1. Data Definition Language:

Used to dfine and describe the contents of the database. Provides a useraccessible data dictionary

2. Data manipulation and query language:

The syntax (language) used to for commands for input, edit, analysis, output, reformatting etc.

Semi standardized eg SQL

3. Data model that supports:

Standardized types: integers, real, character, date

May include specialised data.

4. Most support ultiple views of data.

Users should only see the data they need.

5. security

6. Concurreny control

Manage multi user access to data and contorl updates

7. Database administration tools.

to set the database structure

8. DB design tools

9. Programmable API

allows direct access to the database form other applications

A DATA MODEL REPRESENTS ALL ENTITIES AND RELATIONSHIPS OF INTEREST TO A USER GROUP

a for of metadata that lists the data sets in a system.

Data flow Diagrams (DFDs):

describe flows of data within systems

Entity-Releationship Diagrams ERDs:

Define relationships between database objects.

Unified MOdellingLanguage UML:

graphical and object oriented database modelling.

Data flow diagrams:

provide topdown definition of logical data processes and data flows ina asystem.

emphasize links between data organisation and processes which create, modify, or use data stores.

Entity Rleationship diagrams ERDs

define entities, attributes, and the relationships between entities. this is in comparison to;

DFDs which focus on processes that create,modify or delete data.

can transform an N:M relationship into a 1:M by creating an associate entity.

Proper database design avoids:

storing ireelevent data or omitting required data.

inability ot extend the desing

bad representation of entities

lack of integration with other databases/applicatinos

lack of consistency

'lock in' to a particular format

allow tables to be related together ina temporary view based on common field values.

Overlay

Pycnophylactic

a smooth density funciton, takes into account the effect of adjacent source zones.

It forces local areas to deviate only a litte