Creating Thematic Area Maps With ArcGIS Pro

A thematic map "is used to display the spatial pattern of a theme or attribute" (Slocum et al. 2009, 1). This is distinct from a general-reference map which provides a general overview of information, often representing multiple variables.

A commonly-seen thematic map in the USA is red-state / blue-state choropleth map showing the predominance of political parties in presidential elections.

2012 Presidential Election Results by Party

This tutorial covers the six steps in creating a thematic area map:

  1. Define Your Purpose: The purpose of the map will determine your design choices.
  2. Start the Map: Create a new project or add a map to an existing project.
  3. Add the Data: Add data from the original source.
  4. Symbolize the Data: Create appropriate map symbols from the data.
  5. Label the Data: Add labels to features (optional).
  6. Present the Data: Create a map layout for print and / or publish the map as a service for a web map.
  7. Save Your Project: Save your work as a project package to ArcGIS Online.

Define Your Purpose

When you choose to create a thematic map and are making design choices, you should keep in mind what the purpose of the map is and who the audience is.

Why are you making this map, and what do you want to accomplish?

For example, if you are creating a red-state / blue-state map for a media article, you will probably use the categorical red and blue colors that are generally associated with the parties.

In contrast, if you are creating a map to solicit donations for the campaigns of candidates in evenly-divided "swing" states, you have more freedom to be creative in your advocacy. You might use a continuous diverging two-color scheme that makes the evenly-divided states more obvious than the stark, categorical red-state / blue-state division.

Start the Map

Start a New Project

To create a new map in a new project:

Starting a New Project

Add a Map to an Existing Project

If you have an existing project you can add an additional map to the project by going to the Insert tab and selecting Map.

You may want to rename your maps so it is clear what map is what.

Adding a Map to an Existing Project

Duplicate an Existing Map

If you want to make a map similar to a map you already have in your project, or just want to make a revised version of an existing map without destroying the old one:

Duplicating a Map in an Existing Project

Add the Data

Feature Service

Data sources that wish to make their data available to the public often make it available as feature services. One major provider of feature services in the ESRI ecosystem is ArcGIS Online, which is tightly integrated into ArcGIS Pro. Some services in ArcGIS Online are provided by ESRI, while others are provides by organizations or individuals that use ArcGIS Online to disseminate their data privately or to the general public.

The examples in this tutorial use electoral data from the 2012 and 2016 US presidential election. The data is provided as the Minn 2016 Electoral States feature service from the University of Illinois ArcGIS Online organization.

This data was originally sourced from the Associated Press via Politico and represents election-night returns that do not exactly reflect official final vote totals.

Adding Data From a Feature Service

Data From Shapefiles

The ESRI shapefile is a file format developed by ESRI in the late 1990s. While the format has many limitations and is, arguably, obsolete, it is a well-established format that works with a wide variety of software, so it is still commonly used to share geospatial data.

A shapefile is actually a group of interrelated files that are commonly distributed as a single .zip archive file.

This example uses a shapefile of ward boundaries in Chicago from the city's open data portal.

Adding Data From a Shapefile

Subset the Data (Optional)

There may be situations where you only need to display only some of the features in a layer. For those situations you can use a definition query to select a subset of points based on criteria you specify. You can use multiple clauses to a your definition query if you have multiple conditions.

For this example, we use a definition query to subset only the states adjacent to Illinois.

Isolating a Subset of Features

Symbolize the Data

The choice of how to symbolize the data is based on the characteristics of the data as well as the particular story that you want to tell with your map.

The following subsections describe how to use a variety of different symbologies for different types of variables.

Categorical Choropleth

A choropleth is a type of map where areas are colored based on a single variable that describes some characteristic of those areas. Choropleths can be used to visualize both categorical and quantitative variables.

The following video shows how to create a choropleth using a categorical variable.

Creating a Choropleth With a Categorical Variable

Single-Color Quantitative Choropleths

Choropleths can also be used to visualize quantitative variables. When displaying a single variable, it is common to use a sequential color scheme with a range of lightness or saturation of a single hue that clearly conveys high versus low.

This example uses the percentage of the Democratic vote in the 2012 election. In contrast to the stark, divisively categorical red-state / blue-state maps, this type of map shows that there are Democratic voters in all 50 states.

While this map is not as effective for communicating election results as the red-state / blue-state map (where there is indeed only one winner), this map is more effective at communicating the complexities of the US electorate.

Creating a Single-Color Choropleth With a Quantitative Variable

Two-Color Quantitative Choropleths

There are situations where the purpose of the map is to show divergence above or below a central value. In such cases, the use of two separate colors for high and low values in a diverging color scheme is effective.

An example of this is US election data, where most voters choose between two candidates from two opposing parties. Using the percentage of the Democratic vote by state, red for low values represents more people voting for Republican candidates, while blue for high values represents more people voting for Democratic candidates. The unsaturated grey in the middle indicates a balance.

Like the single-color map, this map offers a nuanced view of the electoral landscape. However, the two-color map also points out balanced "swing" areas where efforts at political persuasion can be effective for winning elections.

Creating a Two-Color Choropleth With a Quantitative Variable

Graduated Symbol Maps

One approach for mapping quantitative values for irregularly sized areas (like states) is to use a graduated symbol map rather than a choropleth. A common example of this is the "bubble" map that uses differently sized circles based on the variable being mapped. Although circles are most common, other types of icons can be used for aesthetic variety.

Graduated symbol maps are also more appropriate than choropleths when mapping counts rather than rates. Because areas rarely have the same populations, creating a choropleth of a count will often just be a map of population rather than a map of the theme you are interested in.

This example is a bubble map of the count of Republican voters in the 2012 election by state. Unlike the red-state / blue-state choropleth, this shows that the large, sparsely-populated states are actually a less significant source of Republican votes than more densely-populated states.

Creating a Graduated Symbol (Bubble) Map

Dot Density Maps

Another approoach for mapping counts is the dot density map, where individual dots represent a certain portion of the overall count.

In this example, this allows us to map the counts of both Republican and Democratic votes simultaneously.

The disadvantage with a dot density map is that dots imply specific locations. Because the dots are distributed randomly across the area, this map does not accurately convey the exact spatial distribution of the voters. This can be remedied by using data for smaller areas (like counties), although data for smaller areas can sometimes be more difficult to acquire and less accurate for sparsely-populated areas where people are difficult to poll.

Creating a Dot-Density Map

Cartograms

Another solution to the irregular area problem is to create a map where the colored polygons are resized and reshaped based on population. This creates significant geographic distortion and is less of a map than a map-like graphic.

This example uses polygons for a continuous cartogram of US states sized by population in the Minn 2020 Cartogram State Continuous layer in ArcGIS Online. This and other cartograms are available here.

Creating a Cartogram

Label the Data

If the geographic areas being depicted should be fairly obvious to most readers, labeling the features can be a distraction. However, you have a limited number of features and you suspect some audience members may not be familiar with the names of the areas depicted by the features, you may want to add labels.

Automatic Labels

Creating Automatic Labels

Annotation Layer

https://pro.arcgis.com/en/pro-app/help/editing/add-or-remove-a-leader-line.htm

One challenge with labels is dealing with situations where labels for features that are close together overlap. ArcGIS Pro tries to move the labels around so they all fit, but the algorithm sometimes doesn't know enough about the data to make appropriate choices. In those cases, you need to create an annotation layer.

Movable Labels in an Annotation Layer

Present the Data

Layout For a Figure

In order to use the map as a figure in a document (such as a report), you need to create a map layout.

Choropleth Print Layout For a Figure

Create a Print Layout For a Standalone Map

If you are going to use the map on its own, such as for a standalone handout or poster, you can add map elements to the layout to format the complete page.

Creating a Standalone Choropleth Print Layout

Save Your Project

Save A Project Package

When you are done with a project, you should save it as a project package on ArcGIS Online:

Saving a Project Package to ArcGIS Online

Reopen a Project Package

Reopening a Project From a Project Package

Lying With Maps

Mark Monmonier's (1991) book How to Lie with Maps details a number of ways that cartographic choices are also choices about what story a map tells. Both the comparative ease of map making facilitated by software like ArcGIS Online and ArcGIS Pro, in tandem with the subtle complexity of standard cartographic techniques, makes it fairly easy to intentionally or unintentionally to tell stories with maps that may not be justified by the underlying data.

The Irregular Polygon Problem

Creating choropleths of areas with widely differing levels of density can overemphasize values. The classic example of this is the red-state / blue-state map that creates the illusion of overwhelming Republican dominance, even when the popular vote majority was Democratic. As described above, this can be mitigated with bubble maps, dot-density maps, or cartograms.

Irregular Polygons vs. Bubbles

The Modifiable Areal Unit Problem

The choice of different types of areas, such as counties vs. states, can alter the results of your analysis of the exact same phenomenon on the ground. For example, the impression given by a red-state / blue-state map is very different than one based on counties or congressional districts. Smaller areas expose fine-grained differences better than than larger areas, but make it harder to see broad patterns.

The Modifiable Areal Unit Problem

Unusual Colors

Nonstandard color choices can create a reversed impression of what the data actually represents. For example, creating a red-state / blue-state map where red is used for Democrats and blue is used for Republicans would be misleading in contemporary America.

Nonstandard Colors

Ill-Fitting Categorization Schemes

Likewise, the use of categorization schemes or thresholds that do not fit the distribution of the values in the data can over- or underemphasize the contrasts between areas.

Differences in Categorization

The Classification Problem

In some cases, using categories rather than continuous color schemes can create clear distinctions when the situation is actually more nuanced. Again, the red-state / blue-state map implies that everyone in a red state is a Republican and vice versa, falling into the ecological fallacy. While such a map is reflective of the binary nature of the electoral college, if you are trying to actually map the political landscape, mapping continuous values emphasizes the purpleness of US society.

Differences in Categorization

Counts vs. Rates

For data like health conditions, mapping counts as choropleths can make the situation in dense, populous areas seem more serious than it may actually be. Normalizing counts into rates, or using graduated symbol or dot-density maps can address this issue.

Counts vs. Rates

Is A Map The Appropriate Visualization?

You should ask if reducing an issue to Cartesian "where" is actually a meaningful representation of the situation. While proximity is still important, advances in communication and transportation mean that distantly separated groups and individuals can have closer social connections than people who are physically closer.

For example, a map of election results by county will show the rural/urban divide, with rural areas leaning toward the GOP and urban areas leaning toward the Democratic Party. However, this requires some understanding of the urban geography of the USA and is inexact. A box-and-whisker plot showing the % GOP vote based on level of urbanization (1 = highest, 6 = lowest) shows much more clearly the dominance of Democrats in urban counties (1) and the GOP in rural counties (6).

The following video shows how to make such a chart.

Creating a Box and Whisker Plot