Department of Geospatial and Space Technology

University of Nairobi

FEB 503: Geoinformation Systems

(GIS)
S. M. Matara
matara@uonbi.ac.ke
GIS Data Modeling
 Representing Geographic Features: Review
How do we describe geographical features?
• by recognizing two types of data:
– Spatial data which describes location (where)
– Attribute data which specifies characteristics at that location
(what, how much, and when)
How do we represent these digitally in a GIS?
• by grouping into layers based on similar characteristics (e.g. hydrography,
elevation, water lines, sewer lines, grocery sales) and using either:
– vector data model (coverage in ARC/INFO, shapefile in ArcView)
– raster data model (GRID or Image in ARC/INFO & ArcView)
• by selecting appropriate data properties for each layer with respect to:
– projection, scale, accuracy, and resolution
How do we incorporate into a computer application system?
• by using a relational Data Base Management System (DBMS)

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 3

 Spatial Data Types
• continuous: elevation, rainfall, ocean salinity
• areas:
– unbounded: landuse, market areas, soils, rock type
– bounded: city/county/state boundaries, parcels
ownership boundaries, zoning
– moving: air masses, animal herds, schools of fish
• networks: roads, transmission lines, streams
• points:
– fixed: wells, street lamps, addresses
– moving: cars, fish, deer

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 4

 Attribute data types
Categorical (name): Numerical
Known difference between values

– nominal – interval
• no inherent ordering • No natural zero
• land use types, county names • can’t say ‘twice as much’
– ordinal • temperature (Celsius or
Fahrenheit)
• inherent order – ratio
• road class; stream class • natural zero
• often coded to numbers eg SSN but • ratios make sense (e.g. twice as
much)
can’t do arithmetic • income, age, rainfall
• may be expressed as integer [whole
number] or floating point [decimal
fraction]

Attribute data tables can contain locational information, such as addresses

or a list of X,Y coordinates. ArcView refers to these as event tables. However,
these must be converted to true spatial data (shape file), for example by
geocoding, before they can be displayed as a map.
21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 5
 Data Modeling
GIS
(Computer System,
Reality
Knowledge Based
How do we model? System)

• transition from reality to computer system involves abstraction process;

thus the model only reflects certain aspects of reality
• only what passes through "glasses of perception" (what fits to the
predefined model) is stored in the computer system
• A model has to represent entities (objects) and their relationships ‘as true
as possible’
• the richness of these models determines the system's range of applications
and capabilities
• specification of the reality to be represented in the GIS – is application
dependent

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 6

 Data Modeling
• There are two fundamental approaches to the
representation of the spatial component of geographic
information:
• Vector Model
• Raster Model Reality
• Data models are the building blocks for the
representation of any spatial information Data model

Polygone: Knoten:

Data structure Polygone: Kanten

P1 l1,l2,l
Knoten arcs
pt 1 l1,l2
P2 l2,l5,l6,l4 pt 2 l2,l5,l6
... ... ... ...

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 7

 Conceptual Model
• A conceptual model describes transformation of reality into given
data structures (a data structure is a particular way of
organizing data in a computer so that it can be used efficiently.)
• objects are structured according to their;
– geometry
– topology
– semantics (thematic information)
Discrete
Geometry
Continuous
Semantics

Object
Relations/Topology

Time

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 8

 Reality

• In the real world we see:

1. Roads
2. Buildings
3. Hills
4. Trees
5. Valleys
6. Rivers etc.

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 9

 Aspects of reality
• Objects
– Phenomena
• E.g. city, river, building
– Facts
• Abstract, e.g. population density, traffic load
– Events, e.g. growth of city
• Properties of objects Semantics
– Width of a river, unemployment rate in city, ...
• Relationship between objects
– Neighborhood, dependencies, size-relations
• Spatial arrangements
– Cluster Relations
– Alignments
• Temporal attributes
– emergency, building of a road, temporal parking Time

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 10

 Conceptual Modeling
• Describe aspects of reality that are relevant for a given task or
application (sometimes derived by system limitations)
• Several aspects should be considered apriori :
– Completeness: modeling should aspire to be as complete as possible
– Accuracy: the quality grade in describing/modeling the object with
regard to it’s reality (in positioning, consistency, etc.)
– Resolution: the size of the smallest object required for modeling (in
raster = pixel size)
• Only what is explicitly modelled can later be used for analysis
and visualization processes
• Questions:
– Which properties do objects have/store/present?
– Which relationships do objects share with other objects?
– To what extent is the modeling essential?

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 11

 Entity-Relationship-Model (ER-Model)
 Example: objects modeled is a lecture hall

# of seats

Projector? Lecture contains

hall

# of floors
Occupation
hours
Is located Address
Building
in

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 12

 Spatial Properties and Relationships
• Length and width (road, river, electricity-line)
• Area (building, city)
• Volume (3D) (building, mountain) A
B
• Perimeter (lake, parcel)
C
• Form (circular, elongated, …)
• Orientation
• Inclination
• Connection (A to B via C)
• Neighborhood (parcels and river)
• Containment (settlements in forest)
• Direction (left, behind, on top)

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 13

 Spatial Distributions
• Regular vs. irregular distributions
• Building in city
• Lakes, islands

• Clustered vs. arbitrarily distributed

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 14

 Spatio-temporal Patterns
• Temporal changes
– Change of land use
– Movement of wildlife
– Change in area size: e.g. growth of a city
– Movement of commuters

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 15

 Thematic Modelling
• layer principle: theme as a function of position
– confer overlay of transparencies with different information, linked by common
geometry (x,y-coordinates/position)
– each layer contains different information
– simple model; easy superimposition
– no hierarchy - all layers are of equal priority
– Each layer has/stores a certain specific geometry
• Advantages:
– very simple model
– map is divided into several thematic layers
– simple overlay by activation of different layers

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 16

 Thematic Modeling
oldest model, commonly used: layer principle

theme 1 layer 1 (geometry and attributes)

Object x,y theme 2 layer 2 (geometry and attributes)

y
theme n layer n (…)
. 317.3 .
Example: Hydrology map
E1 : wells (points) . . 220.1
E2 : streams (lines) x
E3 : rivers (lines)

En : lakes (polygons)

All layers have to be defined in the same coordinate system

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 17
 Representation of reality as Geo-objects
• Modeling with vector data

• Modeling with raster data

• Modeling with triangulated data (vector-based representation)

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 18

 What is Vector Data
• Vector Data uses Points, Lines and
Polygons and their (X,Y) coordinates
to represent spatial features

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 19

 Points
• A point is a 0 dimensional object and
has only the property of location (x,y)
• Points can be used to Model features
such as a well, building, electric pole,
sample location etc.
• Other names for a point are vertex,
node
21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 20
 Lines
• A line is a one-dimensional object that has
the property of length
• Lines can be used to represent roads,
streams, faults, dikes, boundary etc.
• Lines are also called an edge, link, chain,
arc
• In an ArcInfo coverage an arc starts with
a node, has zero or more vertices, and
ends with a node
21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 21
 Polygons
• A polygon is a two-dimensional object with
properties of area and perimeter
• A polygon can represent a city, geologic
formation, dike, lake, river, etc.
• Scale matters

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 22

 Topology
• A set of rules on how objects relate to each other
• The Science of mathematics of relationships
used to validate the geometry of vector
entities, and for operations such as network
tracing and tests of polygon adjacency.
• The study of geometric properties that do
not change when the forms are bent,
stretched or undergo similar geometric
transformations.
21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 23
 Why Topology Matters
• Error Detection
– open polygons
– unlabeled polygons
– slivers
• Network Modeling

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 24

 Raster Data Model

Generic structure of a grid

FEB 503 Geoinformation Systems © Matara

21/09/2017 25
2017/2018
 Representing Data using Raster Model
• area is covered by grid with (usually) equal-sized cells
• location of each cell calculated from origin of grid:
• cells often called pixels (picture elements); raster data often
called image data
• attributes are recorded by assigning each cell a single value
based on the majority feature (attribute) in the cell, such as
land use type.
0 1 2 3 4 5 6 7 8 9
• easy to do overlays/analyses, just by ‘combining’ 0 1 1 1 1 1 4 4 5 5 5
1 1 1 1 1 4 4 5 5 5
1
corresponding cell values: “yield= rainfall + fertilizer” (why 2 1 1 1 1 1 4 4 5 5 5
1 1 1 1 1 4 4 5 5 5
raster is faster, at least for some things) 3
4 1 1 1 1 1 4 4 5 5 5
2 2 2 2 2 2 2 3 3 3
• simple data structure: 5
6 2 2 2 2 2 2 2 3 3 3
2 2 2 2 2 2 2 3 3 3
– directly store each layer as a single table 7
8 2 2 4 4 2 2 2 3 3 3
(basically, each is analagous to a “spreadsheet”) 9 2 2 4 4 2 2 2 3 3 3

– computer data base management system not required

(although many raster GIS systems incorporate them)

FEB 503 Geoinformation Systems © Matara

21/09/2017 26
2017/2018
 Grid Properties
• Each Grid Cell holds one value even if it is
empty.
• A cell can hold an index standing for an
attribute.
• Cell resolution is given as its size on the ground.
• Minimum line width is one cell.
• Rasters are easy to read and write, and easy to
draw on the screen.
FEB 503 Geoinformation Systems © Matara
21/09/2017 27
2017/2018
 Raster Data Model

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 28

 Surface Representation - Points

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 29

 Surface Representation - TIN

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 30

 Surface Representation – 3D - DEM

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 31

 Digital GIS Data
► Real world data exist as:
► Objects - buildings, highways, cities
► Phenomena - terrain, temperature, ethnicity
► Data models for GIS
► Object-based (vector)
► Field-based (raster)

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 32

 Object-based model (vector)
• geographic space is populated by discrete and identifiable objects
• An object
– Has identifiable boundaries or spatial extent
– Is relevant to some intended application
– Is describable by one or more attributes (characteristics)
• Exact objects - are generally man-made features with precise
boundaries
• Inexact objects - are generally natural features with transitional,
or “fuzzy” boundaries

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 33

 Object-based model (vector): representation

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 34

 Points, Lines, Polygons

(3,6)

(4,5)

(7,4.5)
(2,4)

(5,3) (9,3.3)

P(1,1) (9,1)
(6,0.5)

• Data capture: position / centre line / digitization of boundary

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 35
 Field-based model (raster)
• geographic space is populated by one or more
spatial phenomena
• Spatial phenomena
– are real-world features that vary continuously over space
with no obvious or specific extent and are represented as
surfaces
• Surface in a field-based model can be
conceptualized as composed of:
– Grid cells or pixels
• regular tessellations
– Polygons (i.e., triangles)
• irregular tessellations

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 36

 Raster Data
• Gray/RGB values represent
– Light intensity
– Derived values, e.g. land-use classification
– Distribution of noise along a road
– Terrain height Requires image analysis and computer vision
techniques and algorithms (classification)
– ....
row

column
Value Land use
1 Water
2 Settlement
3 Forest
21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 37
 Digital geographic data
• To be usable, digital data must be:
– properly encoded
– properly organized
• Logical organization
– focuses upon data classification and geocoding
– Geocoding is the process of converting addresses (like "1600
Amphitheatre Parkway, Mountain View, CA") into geographic
coordinates (like latitude 37.423021 and longitude -122.083739),
which you can use to place markers on a map, or position the
map.
• Physical organization
– focused upon the way in which the data are stored in the
computer’s memory

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 38

 Geographic Data Precision
• Computer numbers are discrete, whereas real world values are
continuous

• When the original data contain more precise measurements than

those supported by the computer, rounding occurs and precision
is reduced

• GIS coordinates are normally stored as floating-point numbers

(real numbers) in double-precision mode to minimize the impact
of rounding during data processing.

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 39

 Database organization

• attribute (stored field) = one data item

• record (tuple) = group of related items
• data file = collection related records
– ASCII files (alphanumeric);Binary files (0 and 1)
• Digital geographic data files are commonly
referred to as:
– Layers
– Themes
– Coverage
21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 40
 Digital GIS data

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 41

 Raster geographic data representation

• Is best employed to represent geographic

phenomena that are continuous over a large area

• uses tessellations to model a surface

– Tessellations are geometric arrangements (triangular,
square, or hexagonal) of figures that completely
cover a flat surface
– A tessellation is a repeating pattern of polygons
that covers a plane with no gaps or overlaps.

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 42

 Raster versus vector representation
Vector Raster
Advantages  Enables precise point location; thus suitable  Faster data acquisition especially for dense
for representing point data, boundaries and data.
networks  Easier integration of map data with satellite
 Very fine output graphics remotely sensed data sets
 Very suitable for some GIS functions such as  Easier data compaction/compression
network analysis, buffer generation, distance  Suitable for some GIS analysis techniques
and perimeter computation such as map overlay and area computation
 Easier data simulation; each spatial data unit is
the same size and shape
 Some relationships are implicit, hence don’t
need to be specifically encoded

Disadvantages  Data compaction is very difficult.  Results in large data volumes, because when
 Data simulation is not easy as each spatial unit scanning no selection of features is done
is different.  Very unsuitable for representing point features
 Map overlay, a frequently used GIS function  Distances between points are very difficult to
involves very intensive computation establish, hence unsuitable for network and
 Doesn’t allow data integration with satellite boundary data
remotely sensed data, although  Resolution must be very high for fine line
superimposition is possible output, otherwise graphics are coarse.

21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 43

 Choosing a spatial representation
Vector data representation Raster data representation
Focus of model Vector data is focused on modeling discrete features with Raster data is focused on modeling continuous phenomena and images
precise shapes and boundaries of the earth
Sources of data Compiled from:
 Aerial photography  Photographed from airplane
 Collected from GPS receivers.  Imaged from a satellite
 Digitized from map manuscripts  Converted from a triangulation
 Sketched on top of raster display  Rasterized from vector data
 Vectorized from raster data  Scanned blue prints, photographs
 Contours from triangulation
 Reduced from survey field data
 Imported from CAD drawings
Spatial storage  Points stored as x,y co-ordinates  From a co-ordinate in the lower left corner of the raster and cell
 Lines are stored as paths of connected x,y co- height and width, each cell is located by its row and column
ordinates position
 Polygons are stored as closed paths.
Feature representation  Points represent small features  Point features are represented by a single cell
 Lines represent features with a length but small  Line features are represented by a series of adjacent cells with
widths common value
 Polygons represent features that span an area  Polygon features are represented by a region of cells with
common value.
Topological  Line topology keeps track of which lines are  Neighbouring cells can be quickly located by incrementing and
associations connected to a node. decrementing row and column values
 Polygon topology keeps track of which polygons are
to the right and left sides of a line
Geographic analysis  Topological map overlay  Spatial coincidence
 Buffer generation and proximity  Proximity
 Polygon dissolve and overlay  Surface analysis
 Spatial and logical query  Dispersion
 Address geocoding  Least-cost path
 Network analysis
Cartographic output  Vector data is best for drawing the precise shape and  Raster data is best for presenting images and continuous features
position of features. with gradually varying attributes
 It is not suited for continuous phenomena or features  It is not generally well suited for drawing point and line features
with indistinct boundaries
21/09/2017 FEB 503 Geoinformation Systems © Matara 2017/2018 44
 FEB 503 Geoinformation Systems © Matara
21/09/2017 45
2017/2018
 FEB 503 Geoinformation Systems © Matara
21/09/2017 46
2017/2018
 FEB 503 Geoinformation Systems © Matara
21/09/2017 47
2017/2018
 FEB 503 Geoinformation Systems © Matara
21/09/2017 48
2017/2018
 FEB 503 Geoinformation Systems © Matara
21/09/2017 49
2017/2018
 FEB 503 Geoinformation Systems © Matara
21/09/2017 50
2017/2018
 Resulting TIN

FEB 503 Geoinformation Systems © Matara

21/09/2017 51
2017/2018
 FEB 503 Geoinformation Systems © Matara
21/09/2017 52
2017/2018
 FEB 503 Geoinformation Systems © Matara
21/09/2017 53
2017/2018
 FEB 503 Geoinformation Systems © Matara
21/09/2017 54
2017/2018
 FEB 503 Geoinformation Systems © Matara
21/09/2017 55
2017/2018
 FEB 503 Geoinformation Systems © Matara
21/09/2017 56
2017/2018
 FEB 503 Geoinformation Systems © Matara
21/09/2017 57
2017/2018
 FEB 503 Geoinformation Systems © Matara
21/09/2017 58
2017/2018
 FEB 503 Geoinformation Systems © Matara
21/09/2017 59
2017/2018
 FEB 503 Geoinformation Systems © Matara
21/09/2017 60
2017/2018