Data Classification

Cartographic Skills - Mod 6

This week's lesson was on Data Classification.  We learned about several methods of classification but focused on only four for our assignment.  Our assignment was to create two maps, each containing four data frames depicting the same information, displayed by using four different data classification methods; Natural Breaks, Equal Interval, Quantile, and Standard Deviation.  

Both maps were of the distribution of the senior population in Miami Dade County, Fl.  The first map was by percentage of population over 65 and the second was of the actual population of seniors 65 and up by square mile.  This is my map of the population of seniors 65 and up by square mile:

This map is of the distribution of the population of seniors
 65 and up by square mile in Miami Dade County, FL.

I like the map by actual population by area better because I feel it is more honest.  By percentage doesn't tell you how many seniors are there, just what percentage of the population in a certain location is seniors compared to the rest of the population.  If 75% of the population is seniors you don't know if that means three of the four people in that area are seniors or three hundred thousand of four hundred thousand.  With the population by area you at least have a range of actual numbers to go by.

Of the four classifications I don't understand Standard Deviation at all.  Standard Deviation classes are formed by adding or subtracting the standard deviation from the mean of your dataset. Data needs to be normally distributed to use this method and if it is it will give clear dividing points for the classes.  It groups data that is similar and avoids grouping data that is not.  Identical values cannot fall into two different classes.  By the definition of this method it sounds like the perfect method, until you see the results.  The way the data is presented requires at least a basic understanding of statistics.  The color scheme makes no sense until to you compare it to the legend, but when you look at the legend, without that understanding of statistics it makes no sense either.

Equal Interval classes data in equal ranges by dividing the total data range by the number of classes.  Each class represents an equal amount of data, but can be unequal in its distribution.  This method works best on data that is generally spread across the entire range.  It is used by many institutions and government agencies to map complex spatial patterns.  These maps are easy to understand and interpret, but may contain empty classes. It appears there are two, maybe three empty classes in this map.

Quantile classifies data into a certain number of categories with an equal number of values in each category.  This is done by dividing the total number by the number of classes.  Class assignment is based on rank order.  With the quantile method there are no empty classes, but  single class could contain distinctly unlike data values.   

Natural Break method minimizes the difference between data values in the same class and maximizes the difference between the classes by considering the natural groups that are within the dataset.

This assignment was pretty easy as far as the mapping portion went.  The analysis portion was a different story.  I think never having had a statistics class makes a lot of this difficult to understand.  I do feel I have learned a lot from this assignment though and hope I can solidify what I have learned before it all goes away.