GIS Data Models

GIS Data Models

Representing the “real world” in a data model has been a challenge for GIS since their inception in the 1960s. A GIS data model enables a computer to represent real geographical elements as graphical elements. Two representational models are dominant; raster (grid-based) and vector (line-based):

  • Raster. Based on a cellular organization that divides space into a series of units. Each unit is generally similar in size to another. Grid cells are the most common raster representation. Features are divided into cellular arrays and a coordinate (X,Y) is assigned to each cell, as well as a value. This allows for registration with a geographic reference system. A raster representation also relies on tessellation: geometric shapes that can completely cover an area. Although many shapes are possible (e.g. triangles and hexagons), the square is the most commonly used. Resolution is an important concern in raster representations. For a small grid, the resolution is coarse but the required storage space is limited. For a large grid, the resolution is fine, but at the expense of much larger storage space. In the above figure, the real world (shown as an aerial photograph) is simplified as a grid where the color of each cell relates to an entity such as a road, highway, or river.
  • Vector. The concept assumes that space is continuous, rather than discrete, which gives an infinite (in theory) set of coordinates. A vector representation is composed of three main elements: points, lines, and polygons. Points are spatial objects with no area but can have attached attributes since they are a single set of coordinates (X and Y) in a coordinate space. Lines are spatial objects made up of connected points (nodes) that have no width. Polygons are closed areas that can be made up of a circuit of line segments. In the above figure, the real world is represented by a series of lines (roads and highways) and one polygon (the river). A real-world entity could be represented by different types of vector features depending on the map scale used in an application (e.g. a road can be represented as a line at a smaller scale or as a polygon at a larger scale.)