8/14/2019 Data models and management
1/70
DATA MODELS&
MANGEMENT- I
8/14/2019 Data models and management
2/70
Outlines
Introduction
Raster Data
Vector Data
Raster and Vector Structures
Raster and Vector Advantages andDisadvantages
8/14/2019 Data models and management
3/70
Introduction
Geographic Data and Information arethe heart of GIS.
Two fundamental components ofgeographic data: space (expressed asspatial data) and qualities (attributes).
Both of these are stored in database.
8/14/2019 Data models and management
4/70
Data and InformationDefinitions
Information is the primary purpose of
GIS, not just data.
Data is the input; information is theoutput.
8/14/2019 Data models and management
5/70
5
Types of data
Maps
Images
Spatial non-spatial
Postcodes/ZIP codes
Oblique photographs
Videography
Financial statements
Films
Schematic diagrams
KT1 2EE
RH8 9AA
SW1P 3AD12,000 23.45 56789
23,456 12.45 23456
45,987 29.57 87634
8/14/2019 Data models and management
6/70
8/14/2019 Data models and management
7/70
Introduction
Spatial data in GIS has two primary dataformats: raster and vector.
Raster uses a grid cell structure, whereasvector is more like a drawn map.
8/14/2019 Data models and management
8/70
Spatial Data: Vector format
PointPoint- a pair of x and y coordinates(x1,y1)
LineLine - a sequence of points
PolygonPolygon - a closed set of lines
Node
vertex
Vector data are defined spatially:
8/14/2019 Data models and management
9/70
Raster and Vector Data
PointPoint
LineLine
PolygonPolygon
VectorVector RasterRaster
Raster data are described by a cell grid, one value per ce
Zone of cells
8/14/2019 Data models and management
10/70
Raster and Vector Data
Vector format has points, lines, polygons that appearnormal, much like a map.
Raster format generalizes the scene into a grid ofcells, each with a code to indicate the feature beingdepicted. The cell is the minimum mapping unit.
Raster has generalized reality: all of the features inthe cell area are reduced to a single cell identity.
8/14/2019 Data models and management
11/70
8/14/2019 Data models and management
12/70
Raster and Vector DataModels
Raster: because the raster cells value or coderepresents all of the features within the grid, it doesnot maintain true size, shape, or location for
individual features. Even where nothing exists (nodata), the cells must be coded.
Vector: vectors are data elements describingposition and direction. In GIS, vector is the map-like
drawing of features, without the generalizing effectof a raster grid. Therefore, shape is better retained.Vector is much more spatially accurate than theraster format.
8/14/2019 Data models and management
13/70
8/14/2019 Data models and management
14/70
Raster Data
Raster Coding
Resolution Gridding and Linear Features
Raster Precision and Accuracy
8/14/2019 Data models and management
15/70
Raster Coding
In the data entry process, maps can be digitizedor scanned at a selected cell size and each cellassigned a code or value.
The cell size can be adjusted according to thegrid structure or by ground units, also termedresolution.
There are three basic and one advanced schemefor assigning cell codes.
8/14/2019 Data models and management
16/70
Raster Coding
Presence/Absence: is the most basic method and to record afeature if some of it occurs in the cell space.
Cell Center: involves reading only the center of the cell and
assigning the code accordingly. Not good for points or lines.
Dominant Area: to assign the cell code to the feature with thelargest (dominant) share of the cell. This is suitable primarilyfor polygons.
Percent Coverage: a more advanced method. To separateeach feature for coding into individual themes and then assignvalues that show its percent cover in each cell.
8/14/2019 Data models and management
17/70
8/14/2019 Data models and management
18/70
Raster Coding Problems
Raster coding produces spatialinaccuracies.
8/14/2019 Data models and management
19/70
8/14/2019 Data models and management
20/70
Raster Coding Problems
One possible solution is to increase theresolution by increasing the number ofcells, making each one smaller and
therefore more sensitive to accurateclassification.
8/14/2019 Data models and management
21/70
Raster Mapping
A major problem with the raster structure is thatthe shape of features is forced into an artificial grid
cell format.
For right-angled features, such as squareagricultural fields or rectangular political districts,this may not present a major problem. However,for many features, size and shape can becomeundesirably distorted.
8/14/2019 Data models and management
22/70
8/14/2019 Data models and management
23/70
Resolution
Increasing the number of cells on a data setincreases spatial resolution, which helps to
increase spatial accuracy.
One advantage to using relatively few cellsis the short processing time and ease ofanalysis.
8/14/2019 Data models and management
24/70
8/14/2019 Data models and management
25/70
Gridding and LinearFeatures
Low-resolution raster results in a rathergeneralized and crude shape.
High-resolution raster shape appears morerealistic, though still a long way from thevector shape and spatial accuracy.
8/14/2019 Data models and management
26/70
8/14/2019 Data models and management
27/70
Raster Precision andAccuracy
Questions of raster data precision (the exact location) andaccuracy (maximum spatial truth) are often a problem.
Because the raster cell is the maximum resolution and the
minimum mapping unit, there is no way to know exactly wheresmall feature occurs.
Smaller cells have less spatial error because the area of doubt issmaller.
Uncertainty becomes greater when measuring across cells.
Area measurement are also generalized.
8/14/2019 Data models and management
28/70
8/14/2019 Data models and management
29/70
Vector Data
Vector features appear more realistic thanraster features and have better spatial
accuracy.
Vector features are defined primarily by theirshapes, more specifically by the outline oftheir shapes. In GIS, the vector system is acoordinate-based data structure.
8/14/2019 Data models and management
30/70
Vector Data
Shape points are the ends and bends that define the featuresoutline.
At the beginning and end of every line or polygon feature is a
node.
At each bend (change of direction) is a vertex (plural: vertices).
Node are end points and vertices are between, defining the
shape.
Point features are standalone nodes.
8/14/2019 Data models and management
31/70
Vector Data
Chains connect the shape points to draw the features outline.
Chains are vectors or data structure paths that are not part ofthe actual stored data elements; they are not real lines, but
define and present the connection between shape points.
Vector system data files store only the coordinate of each nodeand vertex; the hardware draws the connecting chainsegments. It is virtual component.
The vector data structure is also known as an arc-node modelbecause it uses chains (arcs) and end points (nodes).
8/14/2019 Data models and management
32/70
8/14/2019 Data models and management
33/70
Raster and VectorStructures
Raster and vector structure have differentmethods of storing and displaying spatial
data.
Raster cells are stored and displayed ascells, but in the vector format only the nodesand vertices are stored. This results inconsiderable data storage differences.
8/14/2019 Data models and management
34/70
Raster and VectorStructures
A point in a raster system is a single cell, but in a vectorsystem it is only a node represented by a symbol with itscoordinate position noted.
A simple line in a raster system consists of a sequence ofcells. In a vector system, a simple line consists of two nodesand a chain that connects them.
A more complex raster line consists of connected cells,
sometimes in stair-step fashion when they are diagonal.Complex lines in the vector format have vertices to markchanges in direction, with nodes at each end.
8/14/2019 Data models and management
35/70
Raster and VectorStructures
Raster polygons are filled with cells. Forsingle polygons, the vector format usuallyhas a single node and several vertices tomark the boundary direction changes.
Connected polygons are simply two blocks
of cells in the raster format, but in vectorthey share a common border and somecommon nodes.
8/14/2019 Data models and management
36/70
8/14/2019 Data models and management
37/70
Raster to Vector Conversion
There are at least four basic reasons to convert fromraster to vector:
(1) better visual appearance of vector features;
(2) some plotter work only on vector data;
(3) comparison with vector data is best when bothdata files have identical formats;
(4) some GIS systems have vectors as the centraloperating data structure.
Rasterization of vector data is often called gridding.
8/14/2019 Data models and management
38/70
8/14/2019 Data models and management
39/70
Raster Advantages
A relatively simple data structure;
The simple grid structure makes analysis easier.
The computer platform can be low tech and inexpensive.
Remote sensing imagery is typically obtained in raster format.
Modeling is the creation of a generalized data file or a set ofuniversal procedures to accomplish a certain GIS task.
8/14/2019 Data models and management
40/70
8/14/2019 Data models and management
41/70
Raster Disadvantages
Spatial inaccuracies
Because each cell tends to generalize a landscape, the result
is relatively low resolution compared to the vector format.
Because of spatial inaccuracies caused by datageneralization, a raster format cannot tell precisely whatexists at a given location.
Each cell must have a code, even where nothing exists.
8/14/2019 Data models and management
42/70
8/14/2019 Data models and management
43/70
Vector Advantages
In general, vector data is more map-like.
Is very high resolution.
The high resolution supports high spatial accuracy.
Vector formats have storage advantages.
The general public usually understands what is shown on
vector maps.
Vector data can be topological.
8/14/2019 Data models and management
44/70
8/14/2019 Data models and management
45/70
Vector Disadvantages
May be more difficult to manage than raster formats.
Require more powerful, high-tech machines.
The use of better computers, increased management needs,and other considerations often make the vector format moreexpensive.
Learning the technical aspects of vector system is moredifficult than understanding the simplicity of the rasterformat, particularly when topology is introduced.
8/14/2019 Data models and management
46/70
8/14/2019 Data models and management
47/70
GIS Data Characteristics
Location, or position, is a major staring point ofspatial measurement. Location can be descriptive,or uses a Lat-Lon system.
Size characteristics: Polygon: area and perimeter;Lines: length.
Shape: an important descriptive element used inmap and image interpretation. The shape of afeature often indicates its identity and role on thelandscape.
8/14/2019 Data models and management
48/70
Point features have no real shape or spatialdimension, only the position of objects oroccurrences. They are represented by symbols,such as dots, geometric shapes, or icons.
A line feature has length from beginning to end.
Polygon features have a wide variety of shapes,from easily interpreted circles and squares tocomplicated shapes that defy description.
GIS DataCharacteristics
8/14/2019 Data models and management
49/70
8/14/2019 Data models and management
50/70
Spatial Data Relationships
Spatial relationships are how featuresrelate to each other in space.
It includes distance, distribution,density, and pattern.
8/14/2019 Data models and management
51/70
Distance from one feature to another is anelementary but important relationship. It is
available through simple measurement.
Distribution is the collective location of features;the geographic dispersal or range. There are two
basic ways of perceiving distribution: featuresamong themselves and their spatial relationshipwith other features.
Spatial Data Relationships
8/14/2019 Data models and management
52/70
Spatial Data Relationships
Density is the number of items per unitarea; how close features are to each other.
Pattern is the consistent arrangement offeatures, similar to (and can include)
distribution and density.
8/14/2019 Data models and management
53/70
8/14/2019 Data models and management
54/70
The Data ModelThe Data Model
Geographical variation in the real world isinfinitely complex. Therefore, we require a setof rules (the data model) to convert realgeographical variation into discrete objects.
A set of guidelines for the representation ofthe logical organisation of the data in adatabase (consisting) of named logical unitsof data and the relationships between them.
The GIS Model: example
8/14/2019 Data models and management
55/70
The GIS Model: example
roads
hydrology
topography
Here we have three layers or themes:
--roads,
--hydrology (water),
--topography (land elevation)
They can be related because precise geographic
coordinates are recorded for each theme.
longitude
latitu
de
longitude
longitude
latitu
de
latit
ud
e
Layers are comprised of two data types
Spatial data which describes location (where)Attribute data specifing what, how much,when
Layers may be represented in two ways:in vectorformat as points and linesin raster(or image) format as pixels
All geographic data has 4 properties:
projection, scale, accuracy and resolution
8/14/2019 Data models and management
56/70
Types of data model
The Raster Model
Equivalent of a
continuous grid coveringthe surface, wherebyeach cell in the gridrepresents a square onthe ground.
The Vector Model
Attempts to represent
objects as exactly andprecisely as possible bystoring points, lines(arcs) and polygons(areas) in a continuous
co-ordinate space
8/14/2019 Data models and management
57/70
Raster-Vector Data Model
Raster
Vector
Real World
R ti D t ith R t d
8/14/2019 Data models and management
58/70
Representing Data with RasterandVectorModels
Raster Model area is covered by grid with (usually) equal-
sized, square cells
attributes are recorded by assigning each cell asingle value based on the majority feature(attribute) in the cell, such as land use type.
Image data is a special case of raster data inwhich the attribute is a reflectance value
from the geomagnetic spectrum cells in image data often calledpixels
(picture elements)
R ti D t ith R t d
8/14/2019 Data models and management
59/70
Vector Model
The fundamental concept of vector GIS is that allgeographic features in the real work can berepresented either as:
points or dots (nodes): trees, poles, fireplugs, airports, cities
lines (arcs): streams, streets, sewers,
areas (polygons): land parcels, cities,counties, forest, rock type
Representing Data with RasterandVectorModels
Vector and Raster Models in
8/14/2019 Data models and management
60/70
Vector and Raster Models inGIS
Representation of
Lines
Raster
Vector
TOPOLOGY (for vector
8/14/2019 Data models and management
61/70
TOPOLOGY(for vectordata) What is topology? Why is important? Three types of topological models in GIS Spatial operations of topology
Contiguity Connectivity
Trade-offs of topological structure Application model
Triangular Irregular Network (TIN):Vector-basedGIS
S ti l f t d ti l
8/14/2019 Data models and management
62/70
Spatial features and spatialrelationships
Spatial features in maps Points, lines and polygons
Human being interprets additional
information from maps about thespatial relationships betweenfeatures A route trace from an airport to a house
Land contiguity adjacent to streets alongwhich the lands are located
8/14/2019 Data models and management
63/70
The definition of Topology
The spatial relationships can be interpreted identification of connecting lines along a path definition of the areas enclosed within these
lines
identification of contiguous areas In digital maps, these relationships are
depicted using Topology Topology =A mathematical procedure for
explicitly defining spatial relationship Topology is the description of how the spatial
objects are related with spatial meaning
8/14/2019 Data models and management
64/70
Topological data models
Three types of topological concepts Arc, Node and polygon topologies
Arc Arcs have directions and left and right polygons
(=contiguity) Node
Nodes link arcs with start and end nodes(=connectivity)
Polygon Arcs that connect to surround an area define a
polygon (=area definition)
8/14/2019 Data models and management
65/70
Terms and concepts
To
Node
Left
Polygon
Right
Polygon
From
Node
Connectivity - from and to nodes
Contiguity - Polygon Enclosure
Adjacency - from Direction
Ar
c
Spatial operations of
8/14/2019 Data models and management
66/70
Spatial operations oftopology Connectivity and contiguity (Aronoff, 1989)
A basic, but core spatial analysis operations in GIS Contiguity
A biologist might be interested in the habitats thatoccur next to each other
A city planner might be interested in zoning conflictssuch as industrial zones bordering recreation areas
Connectivity Transportation network, telecommunication systems,
river systems
To find optimum routings or most efficient deliveryroutes or the fastest travel route
To predict loading at critical points in a river channel To estimate water flow at a bridge crossing that will
result from heavy flood
8/14/2019 Data models and management
67/70
Trade-offs of topology
Advantages Spatial data is stored more efficiently Analysis process faster and efficient for large
data sets
By topological relationships, we can performspatial analysis functions, Modelling flow through the connection of lines in
a network (i.e. buffering) Combining adjacent polygons with similar
characteristics (i.e. spatial merge) Overlaying geographical features (i.e. spatial
overlay)
8/14/2019 Data models and management
68/70
Disadvantages
Extra cost and time creating topological structure does impose a
cost Topology should be always updated when a new
map or existing map is updated
Additional batch job working To avoid the extra efforts, GIS systems need to
run a batch job (i.e. a process that can be runwithout user interactions); 70% of total GIS
costs Autoexec.bat in DOS Macro languages such as AML (Arc/Info), Avenue
(ArcView), MapBasic (MapInfo) and etc
8/14/2019 Data models and management
69/70
Conclusions of topology
When topology is created, we canidentify
Know its positions of spatial featuresKnowwhat is around it
Understandits geographicalcharacteristics by virtue of recognising
its surroundings
Know how to get from A to B
8/14/2019 Data models and management
70/70
Thank You