Case Study 4.1 (M)

Geospatial Data & Geodatabase •

LECTURE 2

Professor, Dr. Sergei Andronikov

How to use spatial data to get information • Spatial analysis is MORE than asking questions !

• IN THE CASE OF ENTITIES data analysis concerns

• the attributes, location and connectivity of the entities, and measures of the way they are distributed in space

• IN THE CASE OF CONTINUOUS FIELDS, data analysis concerns the spatial properties of data

• The matter is more complicated by the fact that continuous fields are usually discretized to a regular grid. And the grid can be treated as an individual entity

• DBMS - computer programs for organizing and managing the database. • They use any of, or a combination of database structures. • The aim of DBMS - to make data QUICKLY available to many users.

DBMS: • 1. Allow storage and retrieval of data and data selection;

• 2. Standardize access to data;

• 3. Provide interface between database and application program;

• 4.Allow several users to access the data simultaneously;

• 5. Protect the database from illegal changes.

Most DBMS allow access to data through a high-level programming language and SQL.

Database Management

• GDB – a repository of your spatial data inside a Relational DMBS. • Contains all of your raster, vector data, tables, objects • GDB supports an Object-Oriented Vector data. • Entities are represented as OBJECTS with PROPERTIES, BEHAVIOR, and

RELATIONSHIPS • Object Types include Simple Objects, Geographic features (objects with locations),

networks and topology (with spatial relationship with other features), annotation features, other.

• GDB Model lets you define relationship between objects, together with rules for maintaining their referential integrity.

• The simplest GDB contains a number of independent feature layers (each contains points, lines, areas, annotation)

GDB: Geodatabase format

• GDB stores a feature data itself. User has 2 copies • Advantage – setting up default values for attributes • GDB stores topology, geometric networks, behavior and validation

rules • Feature Datasets contain related feature classes with the same spatial

reference • Domains are maintained in ArcCatalog as a property of GDB rather than of a

single feature class • GDB Annotation is another type of feature class (like point, line) except it

stores labels

GDB: Geodatabase format - 2

• Vectors are well-suited for discrete features. Vector Contents: • OBJECT CLASS. A DB table with which you can associate behavior. “Owners” of “Land Parcels” • FEATURE CLASS. A collection of features. Points, lines, polygons, annotation. Streams,

counties, census tracts. • FEATURE ATTRIBUTES. • SPATIAL REFERENCE • SUBTYPES. A Set of classes for the members of a feature class. • Pipe Feature Class - subtypes: PVC, Iron, Concrete • FEATURE DATASET A collection of feature classes with the same spatial reference. Analogous

to ArcInfo coverages. Vital for facilities networks, roads, environmental layers, census geogr. • RELATIONSHIPS. Association between two objects.

Vector Data in GDB

• GEOMETRIC NETWORKS. A user-defined collection of feature classes that form part of a connected network of edges, junctions, and turns. Water network: valves and meters = junctions, mains and service lines= edges.

• The basic methodology for creating a GN is to determine which feature classes will participate and what role each will play.

• OPTIONALLY – a series of Network weights can be specified. • TWO methods: – A NEW , empty GN – A GN from existing simple features

Vector Data in GBD - 2

• PLANAR TOPOLOGIES. A user-defined collection of feature classes that share geometry. Feature classes as soil types, vegetation, terrain, water can share boundaries. Update all.

• DOMAINS. Define the valid values for attributes as a range or a set of value. TYPES: – Range domain (GPA: from 0 to 4.0) – Coded domain – allows certain values taken from a list (PipeDiam: 1,

3, 6, 12). • VALIDATION RULES. One or more constraints upon the attribute values,

topology or placement of features to enforce the behavioral integrity of features. How features are interconnected in networks. 6-inch & 4-inch pipes.

Vector Data in GBD - 3

• Why GDB? • Enterprise OR Personal GDB

• Large Data holdings: edited, utilized, built

• Choice of creating a Mosaic or a Catalog

• Fast raster dataset display at any scale

• Enhanced raster catalog functionality

• Raster Data compression

• Taking advantage of the Relational DBMS: security, multi-user access, recoverability, etc.

Raster in GDB

• A raster database is built as a number of Cartesian overlays

• 1. In SIMPLE RASTER STRUCTURE where each cell on each overlay is assumed to be an independent unit in the data base, each cell is identified by a coordinate pair and a set of attribute values for each overlay.

• 2. Each overlay represented in a DB as a 2D matrix of points carrying the value of a single attribute. Still requires much storage space.

• 3. The hierarchical structure: many-to-one between attribute values and the set of points in the mapping unit. Each mapping unit is referenced directly. It is clumsy for continuous data.

• 4.Each overlay is stored as a separate file with a general header, and this is followed by a list of values which are ordered according to the sequence of rows and columns. The best one.

Raster Data Structure

Spatial Data • Qualitative and Quantitative data

• Attributes of entities may be expressed by Boolean, nominal, ordinal, integer or real data types (include decimals)

• Differential continuous surfaces require real data types and integers sometimes used as an approximation.

Data modeling & Spatial Analysis • 1. If the location and form of the entity is unchanging, but the attributes can change to reflect

differences then the …………..representation of the entity model is appropriate

• 2. If the attributes are fixed, but the entity could change form or shape but NOT position then - a ……………. model of a continuous field. Drying up lake

• 3. If the attributes can vary, and the entity can change position but not the form, or its parts are linked together, the behavior can be described by an object-oriented model: info from one level to another

• 4. If no clear entities can be discerned then treat it as a discretized, continuous field.

Data modeling & Spatial Analysis • CADASTRE. The main aim is to provide a record of the division and ownership of land.Important:

location, area, extent of land, and its attributes. ………………………………… model works well

• LAND COVER DATABASE. A. The classes are crisp and mutually exclusive, there is a direct relation between the class and its location. Then -…………………………… model, polygon primitives and choropleth map.

• B. But if we use R.S.data and interpolation techniques. Then it is better described by the ……………………………………………… model

• Disciplines concerned with the inventory and recording of static aspects of the landscapes use the entity approach.

• Those dealing with the studies of pattern and dynamic process use the continuous differentiable fields.

Factors to consider • Is the situation/phenomena simple or complex?

• Are the kinds of entities detailed or generalized?

• Do the database entities represent discrete physical things or continuous field?

• Are the attributes obtained by complete enumeration or by sampling?

• Will the database be used for descriptive, administrative or analytical purposes?

• Is the process static or dynamic ?

Develop simple data models for... • A road transport information system.

• The location of fast food restaurants.

• The dispersion of pollution in groundwater

• An emergency unit (police, fire, ambulance).

• A tourist information system.

• The monitoring of vegetation change in upland areas.

• The location of landfill sites and environmental impact assessment study.

• The incidence of landslides in mountainous regions.

• The monitoring of movement of airborne pollution after Chernobyl accident.

GIS in actionGIS in Action