LECTURE 5: VECTOR DATA MODEL

Section I - Review
• Conceptualizing Geographic Entities
As GIS operates on the premise that features in the real world can be represented in a digital
format - referred to as ontology. At the level of an individual geographic entity there are two
possible ways of encoding geospatial data into a GIS.
• Vector Graphics – Commonly referred to line drawings or illustrations where an object
is represented as a point, line, or polygon.

• Raster Graphics – Where an object is represented as a series of pixels (e.g. photo).

Section II – Vector Data Model

• Overview of the Vector Data Model
Vector data models are best suited to geographic entities that have discrete and sharp
boundaries. Additionally, vector data is free to define geographic entity at any spacing in a non-
uniform manner. Check out Figure 4.8 in Campbell and Shin.

Points or nodes reflect specific locations on a map. Some examples of point geographic data
include wells in a county, landmarks, cities on a world map. This is a zero-dimensional feature.

Two points (vertices) connected by a line segment define a line. Examples of lines include:
roads, streams, pipelines, transmission lines. The starting point for a line is referred to as a
node. This is a one-dimensional feature.

Multiple line segments can be connected to form polylines. Polylines that close on themselves
form polygons and include features such as counties, tax parcels, land use, soil types. This is a
two-dimensional feature.

Finally, recall the discussion from last week's lecture of generalization; how features are
represented (points, lines or polygons) depends upon the scale of the map. The importance of
generalization cannot be unstressed and can profoundly influence how a geographic entity is
represented in a GIS. For example, in a small-scale map a city is represented as a point;
whereas, in a large-scale map a polygon is used to represent the cities boundaries.
• A Note on Significant Digits?
Again, the vector data model is best for describing discrete geographic entities that have sharp
boundaries such as streets and buildings. However, the vector model is commonly used to
define entities that have fuzzy and less discrete boundaries. An example of this is a vector
depiction of soil types (Figure 1). Two adjacent soil units are separated by a line that is sharp
and distinct with an inferred accuracy that could be measured in a GIS to many digits pass the
decimal point (for example, 200312.234324 ft). Do all of these numbers pass the decimal point
have real meaning? Can you subdivide these soils along a boundary that is microns thick? No!
You need to write numbers down to the proper level of significant digits.

Decimal Degrees ##.###### o (Six digits pass decimal point)

Planar Coordinates ##.# m or ft (one digit pass the decimal point - unless
you have a really good GPS unit)

Figure 1 (previous page). Soil units 1 and 2, which in a typical GIS have a sharp line as a
boundary but in reality, the boundary between these units is gradational as indicated by the
blue zone between units A and B.
• Is the Vector Data Model Always the Best Choice?
You should realize by now that the stark conceptualization associated with the vector data
model is not always totally valid. Let's go back to the example of our two soil units from Figure
1. This is a classic example of GIS data taking on a life of its own. We now realize that there is
some significant uncertainty in the location of the boundary between these two soil entities (A
and B). The question you may have is then why are we using a vector data model to define
these entities when we know that these boundaries are highly uncertain. Looking at Figure 1 we
can understand that we have two soil units (A and B) with a third transition zone in between
the two units where the characteristics of unit A blend into unit B. So perhaps the selection of a
vector data model is not appropriate. Soils seem to behave more like continuous phenomena as
opposed to a discrete entity. So why not represent these soils in a raster data model? The short
answer is that raster data models typically can only encode one attribute value, whereas, vector
data models can have numerous attributes. This is why soil vector files exist. The user needs to
be mindful that while this data is useful it may not be appropriate for large-scale, highly
detailed mapping applications that occur in the vicinity of soil unit boundaries.

Section III - Vector Data Approaches

• Spaghetti Data Approach
There are two data storage approaches to vector data. An ESRI shapefile is non-topology based.
Another way of stating this is that a spaghetti model is used to store each entity separately as
shown in Figure 2. A way to think of this is to imagine digitizing a pot of spaghetti; each strand is
essentially isolated from each other even if there are touching each other. Figure 2 shows the
issue that can arise in topological data sets. These issues include: a dangling arc, which
overshoots the intended intersection point and undershoots and gaps where line segment to
not connect together as they should. To convert this data into a topological correct network
problems must be addressed through the editing process.

• Topological Data Approach

An ESRI coverage (and other formats) is topology-based meaning that all the entities are
interconnected as shown in Figure 3. Topology is important if you want to be able to use GIS
data to do more then just visualize data. For example, advanced query features such as routing
and network analysis require that your dataset is topologically sound. Basically. all your streets
are connected with each other and are encoded properly. Even tiny gaps or overlaps can
completely throw off your system in a topologic data set.
Figure 2. Two-line segments in a non- Figure 3. Multiple line segments forming three
topological layer. polygons in a topological correct data set.

A nice discussion of spaghetti versus topological approaches is provided in Campbell and

Shin (Chp. 4.2). Also check out DiBiase Section 4.5

ERSI File Formats

Remember that there are two approaches to defining vector data and ESRI has file formats
for both of these approaches.
* Coverage - Esri's closed, hybrid vector data based on topological relations. This is
a legacy format originally designed for ArcGIS Workstation / ArcInfo two decades
ago.

* Shapefile - without topological relations

Shapefile files consist of a number of files that have spatial data associated
with them some of which include:
.shp stores the feature geometry
.shx maintains the spatial index of the feature geometry
and attribute information as is the case with database file (.dbf)

Warning: Do not copy a .shp file in Windows Explorer. For the file to work all
associated files are required
US Government File Formats (Topological formats)
* Digital Line Graph (DLG) - a USGS format for vector data for topographic maps
* TIGER - Topologically Integrated Geographic Encoding and Referencing - from US
Census

Other File Formats

* Geography Markup Language (GML) - XML based open standard (by OpenGIS) for GIS
data exchange
* Keyhole Markup Language (KML) - XML based open standard (by OpenGIS) for GIS
data exchange. Commonly used in Google Earth.
* AutoCAD DXF - Contour elevation plots in AutoCAD DXF format. Commonly used by
architects and surveyors (spaghetti and non-georeferenced data)

Section IV - Overview of Attribute Data

• Introduction
GIS has an advantage in that it can be linked to database management system (DBMS).
Digital data is stored as files (called tables). A database is a collection of tables. Tables are
generally in tabular form. Rows are known as records corresponding to individual
geographic entities, typically geographic objects represented in a vector format. Columns
reflect attributes associated with each entity, which are referred to as fields or attributes.
Typically, each vector entity (point, line, polygon) can have spatial attributes that define the
geographic location of the entity. Additionally, any number of non-geographic attributes
can be included. Conversely, each record in the attribute table is associated with a
particular map feature. Table 1 illustrates an attribute table that contains data linked to
each geographic entity, in this case Dr. McReynold's scorpions. The data contained in the
attribute table provides additional information about the scorpions observed.

• Attribute Data Formats

Note that the spatial data is stored separately from the database of attributes and the two
sets of files are linked with a Feature ID number (typically the left most attribute or field in
an attribute table). The essence of a database management system (DBMS) is that rows
correspond to a mapped entity; whereas, the columns define the attributes associated with
that entity.

Note that within an attribute table there are both geographic and non-geographic fields of
data. The ID, coordinates, and positional references are geographic in nature.
The attribute table in a vector file there has a great deal of flexibility in the types and
number of non-geographic attributes (or fields) that can be stored and several data format
types can be used.
Number (integer ## and floating number - with .####)
Text (short and long string)
Date
Binary large objects (images, multimedia, video, etc.)

Answer the following question based on examining the attribute table depicted in
Table 1.

Identify all geographic attributes

___________________________________ ___________________________________

Identify all non-geographic attributes and note their formats (text, numeric, …).

____________________________________ ____________________________________

____________________________________ ____________________________________

____________________________________ ____________________________________

____________________________________

What type of vector data is this?

(point, line, or polygon) ___________________________________
 MarkSiz
ID EAST NORTH Date Time eClass Location Height Prey Flag
1 684057.6096 17099954.35 2/16/06 21:11 3 ground 0 0 ADS

2 684060.4225 17099946.56 2/16/06 21:20 2 grass 5 0 ADS

3 684072.9915 17099941.42 2/16/06 21:27 1 guajillo 70 0 ADS

4 684058.1735 17099957.55 2/16/06 21:31 1 ground 0 0 ADS

5 684074.4782 17099975.01 2/16/06 21:34 3 Opuntia 30 0 ADS

6 684054.7805 17099966.53 2/16/06 21:40 2 grass 5 0 ADS

7 684084.4272 17099969.01 2/16/06 21:47 2 grass 5 0 ADS

8 684080.6031 17099973.83 2/16/06 21:49 2 Blackbru 35 0 ADS

9 684057.6661 17100016.79 2/16/06 22:05 3 guajillo 40 0 ADO

10 684079.1438 17100022.74 2/16/06 22:11 2 ground 0 0 ADO

11 684157.2014 17099984.14 2/16/06 22:28 4 trunk 35 0 AD1

12 684158.02 17100000.6 2/16/06 22:32 1 grass 5 0 AD1

13 684139.5889 17099983.66 2/16/06 22:41 3 guajillo 55 0 AD1

14 684133.1305 17099988.38 2/16/06 22:44 2 grass 5 0 AD1

15 684125.1963 17099985.82 2/16/06 22:47 4 grass 5 0 AD1

16 683846.3468 17099853.59 2/21/06 21:26 2 Blackbru 105 1 ABS

deadbra
17 683855.6243 17099843.13 2/21/06 21:32 2 nch 65 0 ABS

18 683860.5047 17099849.79 2/21/06 21:32 2 Blackbru 150 0 ABS

19 683833.843 17099862.47 2/21/06 21:40 3 guajillo 70 0 ABS

20 683851.5445 17099900.26 2/21/06 21:48 3 guajillo 85 0 ABO

desert
21 683847.3921 17099935.59 2/21/06 21:54 3 olive 100 0 ABO

22 683850.9261 17099924.01 2/21/06 22:03 2 Opuntia 15 0 ABO

23 683892.1292 17099862.28 2/21/06 22:16 1 Blackbru 135 0 AB1

24 683879.5843 17099890.89 2/21/06 22:22 3 tasajillo 25 0 AB1

25 683712.7985 17099851.35 2/21/06 22:48 2 olive 45 0 AAO

 Geography

Information

Figure 4. Texas county map layer with attribute table shown. Lubbock county is selected;
shown in blue.

In other words, a GIS links features in the real world with information about those features
as you can see in Figure 4. By selecting a county extensive information becomes available
through an attribute table. The map display can be thought of as the tip of the iceberg.
Most of the data is hidden below the surface in the attribute table. Additionally, you can
select multiple counties based on the value of any attribute. For example, you can set up a
query that will select only counties with a population greater than 200,000 allowing one to
visualize all of the urban area throughout the state. GIS allows the use to connect
geographic data with any type of attribute information.

Question: What type of vector data is utilized to depict the counties in Figure 1?

Points or Lines or Polygons

• Levels of Attribute Data
Additionally, geographers have devised a system for classification of attribute data beyond
just a basic description that you observe in an attribute table. Basically, all attributes can be
either non-numeric or numeric in nature. Furthermore, there are two subtypes within each
of these classifications resulting in a total of four fundamental levels of data.

Non-numeric data that is unranked, lacks hierarchy, is referred to as nominal level

attribute data
An example of nominal data is different types of land cover around Laredo indicated below:
Rangeland
Urban
Agricultural

There is no implied ranking in the above list and therefore this data is classified on a
nominal basis.

If non-numeric data is ranked or has a hierarchy then this reflects ordinal level attribute
data. An example of ordinal data is the different types of roads, which has an implied
hierarchy as indicated below:

Most Traffic Interstate

US Highway
State Highway
County Road (Paved & Unpaved)
Least Traffic Trail

The functionality of a geographic entity is directly related to the hierarchy established by

the ordinal classification scheme. For example, if you wanted to evacuate a region you
would more likely direct the public to take the interstate that can accommodate more
traffic than a small county road.

For numeric attribute data, classification is based on whether the number is tied to an
absolute measure scale. If numeric data is not based on an absolute scale this represents
interval level attribute data. A classic example of an interval scale is the Celsius
temperature - 100o C is not double the temperature as 50o C. Another example of interval
data is year. The year 2000 is not double of the year 1000 as our calendar is based on an
arbitrary, non-absolute, system for measuring time.

If numeric data is based on an absolute scale this represents ratio level attribute data. For
example, age is ratio because 10 years old is double the age as 5 years old. Note that
temperature expressed in Kelvin's is considered ratio level data because the Kelvin scale is
based on a temperature of absolute zero. Note that all mathematical operations can be
performed on ratio level data

Also, note that ratio and interval data are frequently grouped into ordinal level categories
for thematic mapping (Lecture 10 and Assignment 4). Why would you want to do this?
You cannot make a map with hundreds of colors but can make one with five or six that
conveys your message/point. The human eye has limits on the number of colors it can
discern on a map and it is a good idea not to exceed 11 colors. Too many colors will make
your eyes "busy" and it will be difficult to interpret the map's meaning.

Now let's go through Dr. McReynolds attribute table (Table 1) again and classify the
attributes in this table both based on data format and level of data. Do not confuse data
format with level of data - data format is how the data is represented in the attribute table;
whereas, level of data indicates the functionality and utility of how data can be used in GIS
applications.

For the following attributes indicate the level of data represented (nominal, ordinal,
interval, ratio)

MarkSizeClass ____________________________________

Location ____________________________________

Height ____________________________________

Prey ____________________________________
Readings
GIS Commons webpage; Chapters 1, 4.
Campbell, J and Shin, M., 2011, Essentials of Geographic Information Systems. Chp. 4.1.
DiBiase, D., 2014, Nature of Geographic Information Systems. Sections 3.7 to 3.10; 4.3 to
4.5.

Terms
Node Point Polyline Polygon
Vertices Ontology Spaghetti Approach Intersection
Small Gap Undershoot Topologic Approach Dangling Arc
Coverage Shapefile TIGER DLG
Attribute Table DBMS Ratio Interval
Nominal Ordinal Thematic Mapping

Concepts
What is the difference between the file format type and data format type in an attribute
table?
Understand how many significant digits are valid for features in the geographic and planar
coordinate systems
Why would a user select a vector data model for a non-discrete feature like soils or land
cover?
Understand the different levels associated with attributes in a DBMS
What are the advantages of using topological datasets? Any disadvantages?
Know about the common file formats used for vector data

HOMEWORK
1. In your own words describe the difference between spaghetti and topological vector data.

2. Comment on the following statement. In some cases, a vector depiction of an entity can
provide a false sense of accuracy regarding the extent of an object.
3. Despite the obvious limitation of a shapefile, this format to this day remains one of the
most popular vector file types. Since ESRI is the GIS software giant this partially explains
the popularity of this legacy file format. Can you think of another reason for the long lasting
appeal of shapefiles?

4. Vector data consists of all but the following

(a) Points (b) Lines (c) Polylines (d) Pixels (e) Polygons

5. Which of the following expresses the correct number of significant digits associated with
a latitude measurement.

(a) 27.45o (b) 27.4512o (c) 27.451254o (d) 27.45125467o

Examining the data associated with Exercise #3 and answer the following review
questions.

6. Fill out the data below based on examining only your vector data

Layer Name Data Model If Vector Type If Vector

(Pt, Line,
Data Approach
Polygon)
For Data Approach Select either Spaghetti or Topological

7. Examining the data packet provided and answer the following review questions. Format
(text versus number, if number is it integer or decimal) and level (ordinal, nominal, ratio,
interval)
Format Level

Stream Layer (LENGTHKM – Attribute) ______________ ______________

Water_Bodies Layer
(GNIS_Name – Attribute) ______________ ______________

Roads Layer (RTTYP – Attribute) ______________ ______________

States_Layer (REGION – Attribute) ______________ ______________

Hydrologic_Units_Layer
(LOADDATE – Attribute) ______________ ______________

8. In the attribute table sort by clicking on top of the RTTYP (Road Type) column. You want
to find the number of line segments used to represent interstate (I).