Introduction to GIS
What is a Geographic Information System?
An Information System (IS) is computer software designed to collect, store, manage, and distribute information.
While most modern Information Systems can store and process spatial data types (descriptions of geometric shapes),
a Geographic Information System (GIS) places more emphasis on the spatial (location) aspect of information.
Specifically, a GIS is concerned with the location of things on Earth, though a GIS can be configured for any
other planet, body or structure by setting up an appropriate coordinate system. From a user interface point-of-view, a GIS
mostly uses the map
as the starting or focal point from which the user interface branches out into other
specialised interfaces.
Compared to Computer Aided Design (CAD) software, a GIS is mostly used for a map
of an area at least the size of
a city block, while a CAD is more appropriate for designing buildings down to components of machines†.
A CAD drawing is most
often in millimetres while a GIS map usually uses coordinates in metres or lattitude and longitude. CAD drawings are
traditionally flat
(using a cartesian coordinate system) or three-dimensional, while GIS maps use either angular coordinates on a
sphere/spheroid or a flat
(projected) coordinate system that can readily be converted to locations on a sphere/spheroid.
†
There is some overlap between GIS and CAD. Many people use a CAD for datasets that are more suited
to a GIS, and it makes sense especially when creating designs/layouts of roads, buildings, etc., even in 2D. For this reason,
most modern CAD systems can use Earth
coordinate systems. A CAD has better functions for design but a GIS is much
better for storing information.
Among many other things, a GIS enables you to:
- Capture & store spatial entities & their attributes;
- Visualize spatial data to reveal relationships, patterns & trends;
- Share information in a form that is easily understood;
- Manage updates to your database over time;
- Integrate with existing systems.
Features
The basic building block of a GIS is a feature†.
A feature is a computer model of a collection of things in the real world that will, for the purposes of the computer model, be identifiable as a single entity. This entity could be a tangible object (e.g., a river) or intangible concept (e.g., an invisible boundary). The computer model of an entity (the feature) contains spatial information as well as descriptive/relevant information. The spatial component is stored as a geometry and each item of information is called an attribute.
†
The word feature
is unfortunately used a lot in software and technology to describe a specific benefit of a product,
i.e., central locking is a feature
of the car. In GIS, feature is the term for an identifiable entity on a map.
When describing the software itself, alternative words like function
must be used in the GIS world.
Attributes
An attribute is an item of information that is included in the computer model of a thing. The most common type of attribute describes a characteristic of a thing, but an attribute could also describe function, purpose, quality, quantity, measurement, expectation, duration, relationship, or any other type of information that may be of interest.
Attributes are pre-defined for a particular class of things, indicating the types of information that the model is concerned (or might be concerned) with. Some attributes are required for a particular application (a problem that needs to be solved, a task that needs to be completed, or a set of such aims). Some attributes may be useful for future applications and are therefore defined because the cost of collecting and storing the information is considered likely to be less than the (potential) benefit that it provides.
Each feature has a unique number, called a FeatureId. Each feature may have a label (text to display on a map) and a measurement. Point features may have an orientation, line features have lengths and polygons have perimeters as well as surface areas. Areas and perimeters/lengths are automatically calculated and updated by PlanetGIS whenever a geometry changes.
Geometry
The geometry of a feature is one of its attributes and has a specific type as described below.
Vertex & edge
A vertex is technically the point where two edges of a polygon meet, but for simplicity we use the term also for points and polylines, to refer to the start and end of a line as well as a single point. The location of a vertex is defined by a set of coordinates (x, y, z or latitude, longitude, altitude) in two- or three-dimensional space.
The geometry of a feature is defined by an ordered set of one or more vertices that may or may not be connected. Two
connected vertices define an edge: a line segment that indicates all the locations along the shortest path between
the vertices. An edge is also a partitioning (division) of space, such that it represents the place where two
faces
(surfaces on a plane, or sides
) meet. A face represents a left
or right
side of
an edge or the inside
and outside
of a closed boundary.
Point
A point geometry represents a location of zero or infinitesimal size. A feature with a point geometry normally represents (models) the centroid of a real-world object that does have a size (sometimes substantial), but for the purposes of the computer model, the size and shape remains undefined (e.g., a point representing a city).
A point geometry is 0-dimensional, not to be confused with its location which can be in two- or three-dimensional space.
In other words, a point has no length and no area and can be identified in the database by the presence of NULL
(blank, meaning not applicable
) values in their Length
and Area
attributes.
Line
Technically a polyline
or line string
, meaning a series of edges defined by a list of vertices, we normally
use the shortened form, simply line.
A line geometry is 1-dimensional, meaning that it has only length. Line geometries can be identified in the database
by a non-zero Length
attribute and a NULL
(blank, meaning not applicable
) Area
attribute.
Polygon
A polygon is a 2-dimensional plane figure, also known as a bounded surface
. A polygon is defined by its
boundary, which is a closed polyline, also known in the GIS world as a ring
. The boundary is the separation of
locations that are considered to be inside and locations considered to be outside the polygon
(or on and off the surface).
Vertices are ordered such that each edge has the inside of the polygon to its left and the outside to its right. Another way to visualise it, is that an observer walking along the boundary and following the order of the vertices, will always have the inside of the polygon on the left. (This is called the left-hand rule.) This also means that the vertices of a polygon's (outer) boundary will be in counterclockwise order.
A polygon should not be self-intersecting, i.e. no edges should cross (intersect) each other. Crossings
should also not occur at a vertex.
Self-intersecting polygons can occur as the result of computer processing and are not rejected, but will misbehave in various ways.
The problem with such polygons is that the left-hand rule is violated. One consequence (of many) is an invalid area calculation.
Note that the above example does not have a vertex at the point of intersection, but even if it had (thereby avoiding intersecting edges),
the polygon would still violate the left-hand rule unless its vertex list started and ended at that intersection point in order
to avoid a crossing
and keep the vertices in counterclockwise order on both sides.
Polygons may have one or more holes
, also called islands. In order to conform to the left-hand rule, the order
of vertices that define the boundaries of islands will be clockwise. Islands must follow, and be completely inside
the outer boundary. An island's vertices may be on top of (coincident with, or touch
) the outer boundary, but may not have any
vertices outside the outer boundary or have edges that intersect with any edge of the outer boundary. Islands may not overlap
each other, though they may touch
.
Since a polygon geometry is 2-dimensional, it has an area measurement. A value (usually square metres)
will be calculated and stored in the Area
attribute. The total area excludes the area of any islands.
The value calculated for the Length
attribute is the perimiter of the polygon and includes
the perimeter of any islands.
Parts
Geometries can also consist of multiple parts. Parts are separated in the vertex list by breaks.
Multi-points
In PlanetGIS, a multi-point geometry is actually a line geometry with a break between every vertex. Because of this quirk
,
a length will be stored, but will always be 0
since the geometry contains no edges that can add up to produce
a non-zero length measurement.
Real-world things that would best be modeled with multi-points are exceedingly rare. Consider using a multi-point geometry where a set of points will share the same attribute information and do not need to be individually identified, for example a cluster of trees.
Multi-lines
A multi-line geometry is a line geometry with breaks, i.e. multiple parts. Each part must have two or more vertices (at least one edge).
Multi-line geometries are useful for modeling things with lines that have more than 2 end-points or that have breaks without requiring the individual parts to be separately identifiable or carry distinct attributes.
Multi-polygons
A multi-polygon is a collection of two or more polygon geometries that are disjoint, meaning they may not overlap. A multi-polygon is a model of a single thing, in other words, all polygons (and their islands) in a multi-polygon, together describe a single surface.
A polygon geometry that has islands already has multiple parts. Therefore, a multi-polygon is a string (ordered collection) of polygon geometries such that the vertices of each polygon's islands are listed before the start of the next polygon geometry's outer boundary.
Polygons in the same collection may touch
or share an edge, but edges may not otherwise intersect any other edge
in the entire collection.
A polygon may be inside an island of another polygon in the same collection. This models the situation where an
island has a lake, which might have an island, etc. This is called nesting
and there is no limit on the number of levels
(depth) of nesting. As long as the left-hand rule is observed and no edges intersect, a multi-polygon can be used to model
any complex surface for which every location is unambiguously either inside/on or outside/off the surface.
Feature classes
Features are classified into feature classes (e.g. rivers or roads) which may be further classified into subclasses (e.g. paved roads or dirt roads). All features of a feature class have the same attributes and subclasses may add more attributes to those of their parent class. For example, features in a feature class “Rivers” might have an attribute specifying whether it is perennial or not.Summary of Concepts
Feature
A computer model of a real world physical object (river) or invisible concept (border).
Feature type
- Point features model the locations of the centres (or representative locations) of entities
- Line features model the centre lines of linear entities
- Polygon features model the surface extents of entities
- Raster features reference bitmap images on a map
- Point cloud features position laser observations on a map
- Web tiles features add background imagery from external sources to a map
FeatureId
A unique number that identifies a feature within a database.
Attribute
An item of information attached to a feature, or the name and type of each piece of information attached to a class of features.
Feature class
A grouping of features that have the same attributes.
Style
A collection of visual elements (symbols, pens, fills & fonts) used to draw the geometries of features.
Display
A reference to a feature class or a subset of features in a feature class, as well as a reference to a style that specifies the visual appearance of the features.
View
A collection of displays in order to create a map for a particular purpose.