GEOG 301 - intro GIS and Maps

Lab 12 - Classification methods and choropleth map production.

The purpose of this exercise is to introduce you to three common ways of classifying data for representation as a choropleth map. The word "choropleth" comes fromt he Greek words 'choros' meaning place or area of magnitude and 'plethos' meaning graphic. Thus, the choropleth map is used to illustrate the relative magnitude of places. The cartographer will normally represent different magnitudes by the use of colors or shading - having each class represented by a variance in color or grey tone. It is convention and perceptually correct to represent classes with lower magnitudes with lighter shades (or values) than those with higher magnitudes for media with a white background such as a printed map.

The main objective of a choroplethic classification procedure is to reduce the number of individual data values transmitted to the map reader. There is no one best way of devising class intervals and class boundaries for choropleth maps. Simplicity is a major goal. It is also worth considering that the class interval system should include the full range of the data and reflect some logical division of the data array in order to portray both the data distribution and the purpose of the map.

Developing class intervals and class boundaries for conventional choropleth maps remains largely an experimental activity. Each geographical phenomenon being mapped presents a unique set of circumstances. Mapping purposes vary, as does the intended readership. The cartographer must look at each new choropleth map from a fresh perspective, without feeling constrained to follow a given method of classification.

The objective of this lab is to play with classification methods in order to understand the different stories different classifications (and numbers of classes) can tell.

First, copy the data to your disk by copying the entire lab12 subdirectory to your flash drive. In this lab, you will be working with the data in the CFY17 Demographics geodatabase. You will note that it includes both WA and USA data.

OK - fire up ArcGIS Pro. Create a new project in the lab12 folder ON YOUR FLASH DRIVE.

Open a new, blank map. Press the add data button and navigate to the data on your flash drive. In this case, you will go into the geodatabase and add the WA County data. Turn off the background layer.

You should see WA in all its glory.

There is a data file associated with this map. To view it, right click on the filename to the left of the ArcMap window and select Open Attribute Table. Check it out and take a look at the available data. Some of the headings are a bit cryptic - you know as much as I do....

Now, create a layout and make sure you can see all of good ole WA, more or less filling the page.

Right click on the wa-county layer on the right and select "symbology". You should see a buncha stuff show up on the left side of the screen.

We will start with a couple of choropleth maps. start by selecting Graduated Colors under symbology. Depending on what you're mapping (nominal data, select unique values; interval/ratio, select Quantiles), you need to select a mapping scheme. As you will first map, say, population, select quantiles - graduated colors. Also, select the field you want to map, colors, etc. Once you select your field to map, you will notice that the Classification area becomes active. Here you can play with the number of classes. Also, hit the Classify button - a new window pops up that gives you details regarding classification method, number of classes, the data, and the breaks. You will note that, somehow, ArcMap decided on 5 classes with the breaks given in the legend area. Play around a little (draw a bunch of different maps with different data/classifications/#of classes. To change this to, say, greyscale (for printing to the laser printer), click on the color ramp options and change this. To find out more about how the natural breaks thing happened, click the classify button. You will note that the classification type used was natural breaks with 5 classes. This is nothing more than ArcMap's default number of classes - it has no relation to the number of clusters in the dataset and probably is NOT the best classification method. Change the classification types to quantile and equal interval. Note that you can change the number of classes.

Alright... another Pro oddity. To see a decent histogram. Right click on the wa_county layer and view the attribute table. Right click on the column for which you want to see a histogram and select statistics. The histogram pops up and you can then play with the chart properties. Especially messing with the number of bins. You can then use this info to better choose the natural breaks.

OK - you now have enough info to play with classification.


Now - make a quick series of maps - pick any of the WA datasets you wish - but use the same column of data for EVERY map. The objective is to compare classification methods, not different datasets. Print each map as you finish it, as making a new one will overwrite your earlier map (you might consider exporting each map to pdf so you have a copy). Note - these will be printed to the laser printer, so make sure your output looks good in greyscale. Don't forget the usual map stuff... title, legend, neatlines, etc. Be sure I know which classification method you are using (perhaps put that info in the title).

Now, we need to use the natural breaks classifier - to do this, you have to first take a look at the data. Fortunately, you can do this in the histogram tool we discussed earlier. Look at where you would put class breaks. And how many. Write down where these breaks should occur. Now, you've figured out what you think is the best option given the data. Make and print out this map.

Examine all 7 classification maps. Given that you have some knowledge of Washington and the data you have been mapping, writeup the following:

  1. Which map(s) seem to give the best representation of the data? Why?
  2. Which map(s) seem to give the worst representation of the data? Why?
  3. Each of the maps tells a different story - pick two of the maps and explain what story they are telling, how that story is being told, and why you (as a cartographer) might want to tell that particular story. Note - only one of the two maps in questions 1 and 2 may be used for this question. By story, I mean.. figure out who you are working for, who your audience is, and what you want to tell them. Then explain how you told that story and how those maps are telling the story you want!


Finally, let's take a quick peek at data normalization. The goal is to make a map of where the kids are o those aged below 14. But to do that, we need to combine three columns (2017 pop ages 0-4, 5-9, 10-14). To do this, look at the attribute table and click on the add button. Name it kids, type is long integer, and format is numeric (double click). Then hit the save button (up at the top ribbon above your map. Close the fields tab, you should see your kids field at the end with no values. Now, we'll calculate values for it - right click on the field name and select calculate field. Add the three field together by double clicking on the field, then "+", then field....

Make a map of where these kids are. Use the Natural breaks classifier with 5 classes. Print it out. Now, let's see where the counties with a higher percentage of kids are located. In properties/symbology window, you just mapped the value as kids, now under that, normalize by 2017 total population. Look at that map. On the map, make a text box and explain how and why normalization changed the map. Print this map too.


In total, you should now have 7 maps of the same data using different numbers of classes and different classification methods and two maps illustrating how normalization affects maps (one of which answers a question on it). Plus a writeup answering the three questions.

Hand in the 9 maps and the answers (typed) to the three questions.