Data classification is a process in which graduated numeric values are grouped into ranges and each classification range is represented by a shade or color on a color ramp, or a symbol size.
The classification method that you use depends on the data you're using and the information you want to convey on the map.
Natural breaks
Natural breaks classification creates classes based on natural groupings inherent in the data. This is the default classification.
Use the natural breaks classification when you want to emphasize the natural groupings in the data. Do not use natural breaks to compare maps created with different data. For example, use natural breaks to compare the number of crimes in neighborhoods across a city. The crime totals will be grouped so that neighborhoods with similar crime totals are symbolized with the same symbol size.
Equal interval
Equal interval classification divides the range of attribute values into equal-sized subranges.
The equal interval classification emphasizes the amount of an attribute relative to other values. Use equal interval for data that has familiar ranges. For example, use equal intervals to compare the total sales at store branches. If you use four bins, the stores will be divided into 25 percent ranges.
Quantile
Quantile classification divides the attributes into bins with equal numbers of features.
The quantile classification can distort the look of a map by placing similar values in different classes. Use quantile classification for data that is relatively uniform. You can also use quantile classification for visual ranking. For example, use quantile intervals to compare the carbon emissions between countries for a given year. If the dataset includes the emissions from 100 countries and you apply 10 bins, you can distinguish between groups of carbon emitters (10 highest emitters, 10 lowest emitters, and so on) but not within groups.
Standard deviation
Standard deviation classifies a feature based on how much the feature's attributes vary from the mean.
The standard deviation classification works best on datasets that are normally distributed and for analyses in which the mean, or the distance from the mean, is important. For example, use standard deviation and a diverging color ramp to compare the average life expectancy between countries. The countries with the highest and lowest life expectancy will be displayed in different dark shades. The colors will become lighter as the classes move closer to the mean global life expectancy.
Tip:
Try pairing the standard deviation classification with a diverging color ramp for choropleth maps. Diverging color ramps style the upper and lower extremes with dark shades and the mean with a neutral color.
Unclassed
Unclassed displays numeric data on a continuous scale, rather than in discrete classes.
Use the unclassed classification when you want to see gradual changes in the data. For example, use an unclassed color ramp to style average temperature measurements for a given time range taken at regularly placed weather stations. The points will show gradual changes in temperature over the study area.
Manual
Manual classification adds custom class breaks that are appropriate for your data.
Manual classification can be used to create new class breaks or modify the breaks created using a different classification method. For example, you may classify the data using equal intervals, then use manual classification to modify the breaks to round numbers.
Use the manual classification when there are known ranges that must be applied to the data, such as when creating multiple maps with the same bins. For example, use the manual classification to compare the number of vacant houses in neighborhoods across a city over time. You can apply the same bins to both maps so that patterns and comparisons can be made without making false assumptions due to differences in the classification.
Resources
Use the following resources to learn more: