Tips on Classifying Data
This section provides definitions and tips to help you optimize your data classification. The Classification Method and # of classes that you chose for your data can affect the outcome of your analysis.
Classification Method
There are four types of classification methods:
Natural Breaks
Divides data into classes based on natural breaks in the data distribution
Calculates groupings of data values based on data distribution and seeks to reduce variance within groups and maximize variance between groups
Limitation: Class ranges are specific to the data set, so it is difficult to compare a map with another map and to choose the optimum # of classes, especially if the data is evenly distributed
Distinct
The only classification method for fields containing text
Creates a separate class for each unique value in the data set
Limitation: if you have a large number of unique values in your data set, your thematic map will be difficult to interpret due to too many different symbols/colors
Quantile
Divides a data set that divide into "n" groups, containing (to the extent possible) equal numbers of observations
Desirable in that this method produces distinct map patterns and minimizes the impact of skewed data
Highlights changes in the middle distribution values because the intervals are generally wider at the extremes
Limitation: areas with similar values can end up in different classes while areas with very different values can end up in the same class
Equal Interval
Simplest method for classifying data
Divides a set of data values into groups each of which represents the same range of values
Better communicates trends with continuous sets of data. Works well with familiar data ranges such as percentages or temperature.
Limitation: not good for non-normally distributed data sets because you might get a map with many features in 1 or 2 classes and some classes with no features
# of Classes
It is the best practice to select 5 to 7 classes for a step ramp. If you select too many classes for a step ramp, it may be difficult to gain insight from the color style on the map. If you would like to visually differentiate a wide range of values, use a smooth color ramp.