Geometric Interval Classification

Geometrical interval classification is a type of classification scheme for classifying a range of values based on a geometric progression. In this classification scheme, class breaks are based on class intervals that have a geometrical series. This classification method is useful for visualizing data that is not distributed normally, or when the distribution is extremely skewed.

The Geometrical intervals classification is better than quantiles for visualizing prediction surfaces, which often do not have a normal data distribution. Geometric interval works best when the data is spread over a large area and is not well distributed. In population data, for example, it is possible to show a better display and distribution of the data in a more natural way. It is possible to see the difference between the more populated areas to medium and low areas, so you can actually see more distribution in the area selected. This classification shows more variation on the data due to the class breaks that happen at a constant geometric increase from the interval preceding the breaks (could be double 2,4,8… or triple 1,3,9…).

The geometrical interval method is similar to a progression classification (binary, geometric, logarithmic, etc.), but it adds a coefficient. Since this method is really intended to be used as part of a data visualization process, it should be noted that it may not be very useful as a data presentation method unless there is a compelling quantitative reason. Best practices may include using a histogram with the class breaks overlaid to show the map audience what the classes mean relative to the data's distribution.

Other methods of data classification used in GIS include Jenks Natural Breaks, Equal Interval, Defined, Quantile, and Standard Deviation.

This method was originally introduced in the ESRI Geostatistical Analyst extension for ArcInfo and was originally called "Smart Quantiles".