Module 4: Data Classification

 Module 4: Data Classification

Data Classification plays a significant role in what a finished map communicates. What a cartographer intends to communicate is significantly impacted by how data are classified. This lab served as an exercise in comparing how different classification methods affect what was communicated on a map, the specifics of how each classification method categorizes data, and when each method might be most or least appropriate. 

Finished Map

Detailed Explanation

Equal Interval:

Equal Interval divides the data into groups equally spaced along the X- axis. If a data set ranges from 0 to 100, and the data is distributed across 5 intervals, then each interval would have a range of 20. In the case of this data, the equal interval ends up being about 2,639 people aged 65 and older per square mile(65andup/sqmi). This reveals that most census tracts have less than 2639, 65andup/sqmi. This map does a great job of showing that geographically most of the county has relatively low density of people aged 65 and up. 

Quantile

Quantile Data divides the data into groups that have an equal number of values within. This map implies that there is greater variance than there may actually be- it exaggerates the differences between classes. Another issue is that vastly different values are being grouped together. The range of the class above in darkest blue is more than 10,000 ppl/sqmi as compared to roughly 400 for the first three.  

Standard Deviation

Standard Deviation shows the statistically significant differences in the data. Its difficult to explain and of limited use to the general population. It does a good job of representing the data and winds up producing a map very similar to natural breaks. 

Natural Break

Natural breaks look for patterns in the data that “naturally” group the data together. This method maximizes the differences between classes. In a way this is grouping likes with likes. Looking at the histogram, in this particular data it is similar to the Quantile histogram, although much less extreme. Because this data is not evenly distributed, natural breaks classification is a good choice for data visualization.