Geographic Data: Levels of Measurement

Data is typically classified as either qualitative (e.g., land use; political affiliation) or quantitative (e.g., per capita income; temperature)—you likely recall learning about this distinction in earlier courses. The classification of your data as qualitative or quantitative will have significant influence on which visual variables you select to map your data. Color hue, for example, is excellent for qualitative data, while color value suggests order or a sequence and thus is probably a better choice for designing quantitative maps.

Nominal is a common term used to describe qualitative, or categorical data. Land use and land cover maps are popular examples of nominal data. They might show, for example, residential blocks as distinct from parks and green space, but this does not suggest that one is lesser or greater than the other.

Figure 6.4.1: A zoning map using the visual variable color hue.

Credit: hayward-ca.gov

Quantitative data can be further classified as ordinal, interval, or ratio data.
Ordinal data has an order, but cannot be presumed to show differences in magnitude. Sports team rankings, for example, describe which teams are better, but not by how much.

Interval data describes orders of magnitude but has an arbitrary zero point. Credit scores, exam grades, and the hours on a clock are all examples of interval data: the intervals between points in all three of these ranges is equal, and none of them have an absolute zero point. Additionally, you can add or subtract interval values, but you can’t multiply them— 2 o’clock plus 3 hours = 5 o’clock, but you can’t multiply 2 o’clock by 3 hours. The classic example is temperature: 0º Fahrenheit and 0º degrees Celsius both serve as the zero point on their respective scales, but refer to different temperatures and therefore arbitrary.

Ratio data, conversely, has a non-arbitrary zero point. Examples of ratio data include counts of forest fire incidence, and yearly household income (e.g., $50,000 is twice as much as $25,000). Interval and ratio data are often grouped together and classified as numerical data.