A place in history: a guide to using GIS in historical research

CHAPTER 2: THE WORLD AS VIEWED THROUGH A GIS

2.3 Vector systems

In vector systems spatial data are represented by either points, lines, or polygons. A point is represented using a single coordinate pair. A line (or arc or segment) is represented by a string of coordinate pairs giving the start and end point of the line and the coordinates of all points where the line changes direction. A line's start and end points are often referred to as nodes. Polygons are created by completely enclosing an area by one or more lines. A basic polygon data model is shown in Figure 2.2. How lines are connected to create this model is known as topology. Both polygons and lines need topological information. A polygon needs to know the ID numbers of all the line segments that make up its boundaries, while a line segment usually knows which polygons are to its left and its right. For example in Figure 2.2, polygon p3 is bounded by line segments s4, s10, s11, s12, and s2. Line segment s10 has polygon p3 to its left and polygon p4 to its right.

Figure 2.2: Simplified structure of polygons in a vector GIS
The spatial data are created by grouping coordinate pairs into line segments. Polygons are then created by completely enclosing an area with one or more line segments. Modified from Jones 1997, 45.

Topology is essential if polygons are used, as without it boundaries would have to be stored twice: for example, line segment s10 would have to be stored as part of polygon p3's boundary and also as part of polygon p4's. It is important to note that the topological data model means that polygons cannot overlap and that every location in the study area can only belong to one polygon. Topology can also be used to turn a collection of lines into a network. This is usually based on each node knowing which line segments it is connected to, so in other words the nodes represent junctions. This can be extremely useful as it means that transport and other networks can be used, for example, to determine the shortest path between two points.

Figure 2.3: Linking spatial and attribute data
The attribute data are stored in a conventional database table and linked to the spatial data using polygon ID numbers. Point and line data are linked to attributes in the same way. Modified from Jones 1997, 45.

Attribute data is linked to points, lines, and polygons using a relational join on ID numbers as is shown in Figure 2.3. In modern software the creation and maintenance of topology, its associated ID numbers, and the links between the spatial and attribute data are kept hidden from the user. Nevertheless it is important to have some understanding of how the data are assembled.

Some features, such as towns or buildings, can be represented by either a polygon or a point. The choice of which representation to use is dictated by the purpose they are to be used for and, in particular, the scale of use. Polygons can be represented by points if required usually by using the centre point of the polygon, termed the centroid.

The Great Britain Historical GIS (Gregory and Southall 1998; 2000), SECOS (Gatley and Ell 2000) and the Great American History Machine (Miller and Modell 1988) all have attribute databases containing a variety of census and other data. These are linked to polygon representations of the administrative units used to publish the data such as registration districts in England and Wales, and counties in the United States. In this way the data are put on the map and the map is populated with data.