Geographic Information System

Spatial Data Modeling

Lecturer:
Eng. Thomas Matabaro
[email protected]
 Outline
 Geographic phenomena:
 Objects and fields
 Discrete and continuous fields
 Boundaries
 Spatial data model:
 Vector data model
 Raster data model
 Data structure
 Vector data model
 Raster data model
 Geographic phenomena (1)
 Geographic phenomena exist in the real world
 Geographic phenomena are a manifestation of an
entity that can be:
 Named
 Geo-referenced (geographic)
 Assigned a time at which it is or was present

 There are different types of phenomena,

 To recognize them, we can select the correct
way to store them in GIS environment.
 Geographic phenomena (2)

Air temperature
Shoreline  Geographic phenomena are
the studied objects of GIS.
 Geographic phenomena exist
Soil in the real world, everything
you see outside is a
type
Geographic phenomenon.
 Some of the things you do
not see are also Geographic
phenomena like temperature.
Elevation Rocks
Water temperature
Geographic phenomena (3)
 We need to come up with a
digital representations of
geographic phenomena in
tessellation order to store them in a GIS.

 This is not easy because

isolines different phenomena require
different digital
representations

 Also, multiple representations

TIN are possible for the same
phenomenon.
 Types of geographic phenomena (1)
 Two types of geographic phenomena: field and
object
 A field is a geographic phenomenon for which
every point in the study area a value can be
defined.
 An object is geographic phenomena that do not
cover the entire study area,
 The space in between two objects is empty or
defined.
Types of geographic phenomena (2)

 Typical examples of fields:

temperature, barometric pressure
and elevation.
 In picture on the left you see an
example of elevation.
 You can measure the height
(altitude) everywhere.
 Even if an area is flat, elevation
can be measured at any point.
Types of geographic fields (3)

 Two types of geographic fields: discrete and continuous

fields.
 Continuous field:
 The principal function is assumed to be continuous.
 Continuity means that all changes in field values are
gradual (example elevation).
 Discrete fields:
 They cut up the study area in equally exclusive
bounded parts,
 with all locations in one part having the same field
value (for example land use)
Types of geographic fields (4)

 Discrete: Categorical and Discontinuous data

represent objects belonging to:
 A class (soil type),
 A category (Zoning), or
 A group (a district hospital)
 Continuous : Non discrete and surface :
 Usually, it represents the continuity in data like
elevation, slope, radiation levels, pollution levels,
noise… etc
Continuous Fields

 Continuous means that all

changes in field values are
gradual
Slope  In a differentiable field we can
measure the change.
 In the example on the left, we
can measure the gradient
slope as the change of
elevation.
Discrete fields (1)

forest  Discrete fields cut up the

study space in subparts
agriculture
with a clear boundary,
with all locations in one
Valley part having the same
value.
 Typical examples are land
agriculture
classifications, geological
classes, soil types, land
use types, crop types or
natural vegetation types.
agriculture

road
Discrete field (2)

 Objects are discrete and

bounded entities.
 The space between objects is
empty or not defined.
 Here, the space is not really
empty, it my contain roads,
gardens, driveways etc, but
house these are not houses or
house buildings.
No house: empty
 Object
 The position of an object in
space is determined by a
combination of one or more of
the following parameters:
 Location (where is it?)
 Shape (what form?)
 Point
 Line
 Polygon
 Size (how big?)
 Orientation (which direction?)
The river is an object, with a
location, Shape and a direction
 Data types and values
• Different types of values that we can use to
represent “phenomena”.
• Four different data types:
1. Nominal data values
2. Ordinal data values
3. Interval data values
4. Ratio data values
• Nominal and Ordinal data together are often referred
to as ‘qualitative’ data, Interval and Ratio are known
as ‘quantitative’ data
Different kinds of data values
• Nominal or categorical data
– Values provide a name (identity)
• E.g. ‘valley’, ‘deciduous forest’, ‘Paris’
• Ordinal data
– Values can be put in a natural sequence or order
• E.g. income level categories: low-medium-high
• Interval data
– natural sequence of values with equal intervals, but
with arbitrary zero point
• E.g. temperature
• Ratio data
– Values have a natural zero point, allow for computation
• E.g. distance, weight
Boundaries (1)  Both objects and
discrete fields have
boundaries
Crisp
 Two different types of
boundaries:
Boundary
 A crisp boundary is
one that can be
determined with
precision
 Fuzzy boundary is not
Fuzzy a precise line, but
rather an area of
transition.
 Spatial data model
• Allows the geographic features in real world
locations to be digitally represented and stored
in a database so that they can be abstractly
presented in map form, and can also be worked
with and manipulated to address some real
world problem
Data model and Data Structure (1)
Data model:
 Computer representations are divided in two groups:
raster and vector data models.
 The next step is to understand how data models can
be applied to represent geographic fields and
objects.
Data structure:
 Data structures provide information that computer
requires to reconstruct the spatial data model in
digital form.
 Two types of data model

2 4 4
Lac 2 4 4 4 4
1 1 2 4
1 1 1 2
1 1 1 1 2
Forest
1 1 1 1 2
1 1 1 2
1 1 1 2 3
1 2
House
2
2

Vector Data Model Raster Data Model

Vector data model
 Derived from the formulation of spatial concepts that
emphasize on real world objects
 Geometry primitives of vector data model are point, line
and polygon.
 A point feature is a spot (or location) that has no physical
or actual spatial dimensions.
 A point feature represented by a single coordinate and
only has a geometric property of location. A line is a one-
dimensional feature having only length, no width.
 A line is represented by series of points and has the
geometrical properties of location and length.
 Polygon: A spatial feature that is represented by a series
of lines and has the geometric properties of size and
perimeter.
Raster Data Model
Reality  In the raster data model,
individual cells are used as
the building blocks to create
a total map.
 The cells are of the same
shape and size,
Building
 The field attribute value
Road assigned to a cell is
associated with the entire
Field area occupied by the cell.
Raster data model

Origin  Pixel or Cell= a

Pixel (cell)
square representing a
specific portion of an
area. Always with the
same size
 Rows & Columns =
Value
Cartesian matrix.
Line 9
Each cell has a
unique row/column
Column 3 address
 Values = one value by
pixel
 Examples of Raster Data
Satellite Imagery

Classified image

Scanned Image

 Aerial Photography
 Raster Structure  Besides regular
tessellations, there are
Normal raster – regular tessellation also irregular
tessellations:
 Partition of the space
into mutually disjoint
cells.
Quadtree – irregular tessellation  Cells vary in size and
shape
 Adapts to spatial
phenomena
 Example: Quadtree
Vector Data Model (1)

 A vector data model uses X and Y co-ordinates to

store the shape of a spatial entity.
 In the vector model, the point is the basic building
block from which all spatial entities are constructed.
 As such, the simplest spatial entity is the point.
 Lines and polygons are constructed by connecting a
series of points into chains and polygons.
 The more complex the shape, the greater the number
of points needed.
Vector Data Model (2)

 Points are defined as single

coordinate pairs (x, y) when we
work in 2D, or as coordinate
triplets (x, y, z) when we work
in 3D.
 Points are best used to
represent objects that are
described as shape.

Points representing
trees along a road
Vector Data Model (3)

Line representations:
 Used to represent one-
dimensional objects (roads,
railroads, canals, rivers…)
 Line is defined by 2 end nodes
and 0-n internal nodes to
define the shape of the line.
 An internal node or vertex is
Begin node Vertex
Line or arc
like a point that only serves to
define the line.
End node
Vector Data Model (4)

Area representations:
 When area objects are stored
using a vector approach, the
usual technique is to apply a
boundary model.
 The area is defined by de
boundary of the area

You store the

boundary of the area
Vector Data Model (5)

 A simple representation of the

area features would be to list
for each polygon the list of
lines that describes its
boundary.
 This is called a polygon-by-
polygon representation.
 Each line in the list would be a
sequence that starts with a
node and ends with one.
Total boundary of the
polygon
Vector Data Model (6)

 The reason why this is not a

good representation is called
data redundancy.
 This means that shared
boundaries between polygons
are stored double.
When storing the second
boundary , some line
segments are duplicated
Vector Data Model (7)
Line 3

Line 2 Topology
 The boundary model or
topological data model is an
improved representation of
Line 1
the polygon-by-polygon
model.
 It stores parts of a polygon’s
boundary as separate line
segments.
 Vector Data Model (8)

3
4 Line D
Topology
Line E 7

5
Line L
Line K
6
 It also indicated which
Line M ZZ
8
Line J
XX polygon is on the left
and which is on the
Line F
15 9
Line I
10
Line H right of each arc
Line N Line O 12
Line G 11
14 13Line Q
Line P Line From To Node Left Right
QQ Node polygon polygon
N 15 14 QQ

P 14 13 QQ

O 15 13 ZZ QQ

I 9 10 XX ZZ
Vector Data Model (9)

 We can determine the left

and the right polygon,
because the line segment has
a direction.
 The direction of the line
From Node 15 segment is from the “From
Left Node….” to the “To Node….”

Right

To node 13
Topology of Two dimensions Spatial
Relationships
The topology of Two Dimensions Set Theory

• We use the topological properties of interior and

boundary to define relation-ships between spatial
features.
Topology of two Dimensions Set Theory
 Reading Materials

• Otto Huisman and Rolf A. de By (2009): Principles of

Geographic Information Systems, ITC, Enschede, The
Netherlands. Chapter 2.
Thanks for your attention