Thematic Mapping of Percent of Persons claiming Two or More Races at the U.S. Census Bureau American Factfinder Web Site 

by 

Dr. James R. Carter, Geography-Geology Department

Illinois State University, Normal, IL, USA

Start by going to the American FactFinder web site.  In the block on the left, click on Maps and Geography.  That will take you to the Maps in American Factfinder page.  Once there click on Thematic Maps, or click here to get the Thematic Map page.  When you get here you should see a map of the states of the U.S. showing persons per square mile, what we call population density.  There is no title on the map but the subject of the map is given in the legend.  The density values on the map are broken into five classes and symbolized by the green color sequence.  You could click on the word Legend written in blue and change many aspects of this map.  But, because of the great variation in size of the states, and the District of Columbia, a map of population density is not very meaningful.  So, for this exercise we will pick a new variable and map it. 

Look for the Thematic Maps legend and click on Themes.  That is shown here in magenta.

This will bring up a table of Census variables.  Look through these variable and see the nature of the data that is available.  Observe that many of the variables are given as absolute counts--as Persons of Two or More Races.  If you map absolute counts, it is similar to mapping population density.  There are likely to be many more persons of this nature in New York, Chicago and Miami simply because there are many people in those cities. 

For this study I have chosen to look at Percent of Persons of Two or More Races: 2000.  But first, a comment on this variable.  In the United States we are still sensitive to race, although there is great debate as to what we mean by race.  In Census 2000 there was an option for the first time to claim that you were of two or more races.  I like to think of this as the Tiger Woods question, for he openly acknowledges that he has ancestry of many different races.  So, when we look at this variable we are looking at those persons who choose to claim that they were of two or more races. 

Below you see where I toggled down and selected variable TM-P004G Percent of Persons of Two or More Races: 2000.  Then I click on Show Result on the right.

That results in the map below being brought up. This is the default map of the site.  The default map shows the states of the U.S. on the aqua background.  Alaska, Hawaii and Puerto Rico are in separate boxes at the side.  The statistical data are broken into five classes using natural breaks to divide the classes.  There is no title on the map and in the legend it states the data are in percent numbers.  Looking at this map there is no way to tell what variable is being mapped.   

In the legend note that the highest class has the value 21.4% and that only Hawaii is shown by the darkest green.  Yes, Hawaii stands apart from the other states on this variable--way apart.  Now it is time to play with this map and customize it.  Click on the blue Legend.  This should bring up a box showing options to change the map. 

In this case I want to change from the 5 classes of Green to 4 classes using the Orange color sequent.  I will continue to use Natural Breaks.  I also want to insert a Title.  So, click on Title and type in your title but note the title will be flush left and write over a portion of Alaska.  For the title below I added many spaces on the left.  When you have everything you want, click on Update.

With the Orange color sequence, Hawaii is shown in Red and stands out as very distinct.  I return to the Legend and created the map below by selecting 7 classes using the Orange sequence.  This new map is quite different in appearance from the one above.  In the map below it is difficult to see that Hawaii at 21.4% is distinctly different from Alaska, California and Puerto Rico, which fall between 4.2 - 5.4%.

A common way to classify data is to break it into Equal Steps.  The map below breaks the data into 7 Equal Steps, showing a range of about 3.0% within each class.  The result is that most states fall into the lowest class, five into the second class and Hawaii into the top class.  No states fall into the third, fourth, fifth and sixth classes.  Is this bad?  No.  This map shows that there is a great difference between Hawaii and all other states.  It is important to make that point.

The image on the map above is very different from the map below, where the data are broken into seven quantiles--that is one-seventh of the states fall into each category.  On this map Hawaii shares the top category with many other states, but that category is very broad.

The map below was created using data classes that I defined myself.  Under User Defined I selected round numbers as best I could.  On this map the lowest class is 1.0 or less.  The next class is from 1 to 2, then 2 to 3, etc.  I made the fifth class as 4.1 to 5.5 to include the highest states but not Hawaii.  Then no state falls into the sixth class and Hawaii stands alone as the only state in the top class.  Personally, I think it is important to employ many variations of the same map to analyze the patterns.  But, if I had to employ only one map I would probably choose this one.

The maps above are based on State level data.  To map county level data go to the Geography option and select County.

This will bring up a map similar to that below.  In this case I had the 4 classes, natural breaks, Orange sequence when I selected Counties.  The pattern looks quite different at the County level of aggregation.  Observe that the values in the Legend now range from 0.0 to 28.4%.  At the county level of aggregation you should expect to see greater variations.  Because it is hard to see many counties at the scale of the screen, I zoomed in to focus on the northeastern portion of the United States for subsequent discussions. 

Your map may come up showing highways and rivers.  You can go into the Legend and turn these off.  Now look at the map above relative to the map below.  The map of the entire U.S. is on a conic projection as is evident by the curve of the parallels.  It is most obvious on the boundary extending east to west across North Carolina, Tennessee Oklahoma, New Mexico and Arizona.  The map below is on a rectangular projection as shown by the straight lines representing those same boundaries. 

The map above is based on 7 natural breaks.  Note that the range of values is based on all of the counties in the U.S. and not just those shown on this map.  There is no way to change this in this program.  Look at the high places on this map--eastern Oklahoma, northern Michigan and a few scattered counties.  In most cases these are areas with large populations of American Indians.

This map uses the 7 classes and the Orange color sequence but the data are broken in Equal Ranges.  Based on this common classification scheme everything looks the same.  But, this is a valid map because in most places less than 4% of the persons claim to be of Two or More Races.   But, when the data are broken in to 7 quantiles the pattern changes dramatically, as below.  But, look at the range of values in that top category, from 2.6 to 28.4%. 

Using Quantiles to classify data is very common.  We talk about the top quarter or upper 50% or top 10%, or the bottom third, etc.  But, as shown here much can be masked with that classification method. 

If you have gotten here you should be ready to use the American FactFinder web site to explore the nature of choropleth maps (thematic maps) and the nature of Census data.  Do note you can download the Census data behind each map by clicking on the Print/Download option.  You should download the numbers and examine the nature of the distribution.  Having done that for this data I found that only six counties or equivalent have values of 10% or higher. 

With more than 3,219 counties or equivalent, it is not practical to make a graph similar to that used on the previous web page.  But by looking at some extremes, you can get a good feeling for the nature of the data.  Such big differences do not exist at the low end of the scale in this data set. 

If you are going to make thematic maps, you owe it to yourself and to your audience to learn about your data and make maps that are representative of the data.  Hopefully, you appreciate the value of looking at many maps showing the same subject.  You are the person who has the option to pick the number of classes, the classification scheme, select the color sequence and give a title.  Exercise those options.

This page is in support of the course Geography 204 - Geographic Techniques at Illinois State University. 

Return to the master pages of Geo204 

Return to the master pages of James R. Carter, the author of this page