In this chapter, you’ll learn how to design and symbolize thematic maps. A thematic map strives to solve or investigate a problem, such as analyzing access to urgent health care facilities in a region, as you did in chapter 1. A thematic map consists of a subject layer or layers (the theme) placed in spatial context with other layers, such as streets and political boundaries.
Choosing map layers for a thematic map requires answering two straightforward questions:
Quite often, the subjects of thematic maps are vector map layers (points, lines, or polygons), because such layers often have rich quantitative and qualitative attribute data that is essential for analysis. Of course, the subject can be a raster layer (in chapter 10, for example, you will create a risk-index raster map to identify poverty areas of a city, and poverty is the subject of the map). Spatial context layers can be vector, such as streets and political boundaries. These layers also can come in both raster or vector formats, including many basemap layers provided by Esri map services.
The major map design principle for thematic maps is to make the subject prominent while placing spatial context layers in the background. For example, if the subject is a map layer with points and you want to give them focus, you might give the point symbols a black boundary and bright color. These subject features are known as “figure” and are the main composition of the map. Everything that is not figure is known as “ground.” For example, if a context layer has polygons that are not the focus of the map, you might give the polygons a gray boundary and no color, thereby placing them in the background.
Symbolization is easy for vector maps because ArcGIS Pro can use attribute values to automate drawing. For example, ArcGIS Pro could draw all food pantry facilities in a city by using unique values with a square point symbol of a certain size and color. Continuing, the software could draw all soup kitchen facilities with a circle of a certain size and different color by using an attribute with type-of-facility code values (including “food pantry” and “soup kitchen”).
In this chapter, you will learn to use good cartographic (symbolization) principles as you build several vector-based thematic maps.
Placing objects of all kinds into meaningful classes or categories is a major goal of science. Classification in tabular data is accomplished using attributes with codes that have mutually exclusive and exhaustive qualitative values. For example, a code for size could have the values “low,” “medium,” and “high.” Any instance of the features with this code is displayed in only one of the classes (the values are mutually exclusive). Moreover, there are no more size classes (the values are exhaustive). In this tutorial, you learn how to symbolize mapped features—points, lines, and polygons—by class membership as available in code attributes.
The Neighborhoods and Water polygon layers provide spatial context. Such layers should be displayed using outlines with no color fill, with water features being an exception and generally given a blue color and no outline. Context layers are easy to symbolize. You can start with Neighborhoods.
Turn on the Water layer and symbolize the layer with a blue polygon symbol. Hint: On the Gallery tab, search for Water and click one of the Water (area) symbols.
The last layer to symbolize, Zoning Land Use, is the subject of the map, displayed by Unique Values on primary land-use code. Land-use maps use muted colors, which you’ll create next.
Next, you’ll assign colors used by the New York City Planning Department. You’ll start by changing the outlines of all polygons from black to light gray. When you view land-use polygons, their black outlines often take up too much of your map and attention, and also distract from the symbolized color. A gray color will soften this interference and still show boundaries.
Labels created from attributes such as neighborhood names are an important part of cartography and an integral and informative component of a map. You must specify the elements of font, size, color, placement, and visibility ranges to make labels easy to read.
In this exercise, you will label all three layers of the map. Each layer will have its own label properties and label placements.
The field in this map has detailed zoning codes known by developers, planners, and other members of the user community.
Use the Lower Manhattan bookmark. Label the Neighborhoods layer using Name, Arial font, Bold, Size 7, and a white halo. Finally, on the ribbon, on the Labeling tab in the Label Placement group, click Label Placement > Land Parcel.
Label the Water layer using Landname. Use font Times New Roman, Italic, Size 12, and the color Atlantic Blue.
Set the Neighborhoods and Water labels to turn off when zoomed out beyond the Lower Manhattan bookmark. Try out the labels by zooming in and out and using bookmarks.
Labels for some water polygons might overlap with redundant and unnecessary labels. Removing duplicate labels will unclutter the map. You will do so using another menu option to set label properties.
Often, a map layer has more features than you want to display. If so, you can use a definition query to display the desired subset of features from the larger collection, on the basis of values in the feature attribute table. For example, the point features in this tutorial start with point features for all facilities in New York City (food, health care, fire and police, schools, senior centers, and so on). You will want to display the features for food facilities only, including food pantries and soup kitchens. Query definitions allow you to select and display just these features. A definition query is different from the Select by Attributes in chapter 1. The definition query is used to filter the display of a layer rather than selecting a temporary subset of features to work with, even though they both use a similar SQL interface.
In this exercise, you will create a query to display a subset of more than 20,000 government and nonprofit facilities in New York City. Facilities is a point layer of locations for services that the city provides. The map needs only three out of more than 100 classes. Classes have both a numeric code (Facility_T) and a corresponding description (Factype_1), and the three needed facility classes are 4901 = Soup Kitchen, 4902 = Food Pantry, and 4903 = Joint Soup Kitchen and Food Pantry. Showing the location of these facilities could help the directors of New York City’s food banks determine whether they are well located relative to poverty areas of the city.
The subject of the map, Food Facilities, is figure, and all other layers are ground. Figure features get accentuated with bright colors, and ground gets shades of gray.
Next, you’ll symbolize the three types of facilities. In this example, varying shape and color for unique point symbols is good practice. Color-blind people can use shape to identify facility classes; also facilities will remain distinguishable in black-and-white photocopies of such symbolization. The majority of people who can see color get the full effect of shape and color for seeing patterns of food facilities.
A policy decision-maker might want to know the streets where food facilities are located and doesn’t want the details of a basemap. Turn off the World Light Gray basemap, and turn on Manhattan Streets. Display Manhattan Streets as a ground feature using gray (20 percent) with width 0.5 pt. Zoom to a few blocks in Manhattan and experiment with various label properties. Save your project.
Showing continuous variation in numerical attributes is not possible when you use the attributes to symbolize points or polygons on a map. The human eye simply cannot make distinctions unless there are relatively large changes in graphic elements. You must break a numeric attribute up into relatively few classes (roughly three to nine), similar to how you create a bar chart for a numeric attribute. Each class has minimum and maximum attribute values. The minimum value is included in the class, but the maximum goes in the next classification to the right. To symbolize map features, you need only the set of maximum values for classes, called “break points.”
A choropleth map uses color in polygons to represent numeric attribute values. Generally, increasing color value (darkness of a color) in a color scheme represents increasing (higher) values. In this exercise, you will use US Census data aggregated to New York City neighborhoods to create choropleth maps for households with persons over age 60 receiving food stamps/SNAP (Supplemental Nutrition Assistance Program).
Choropleth maps use classification methods to display the data, and methods will vary depending on the data and intent of the map. The default classification method is Natural Breaks (Jenks). This method uses an algorithm to cluster values of the numeric attribute into groups, with the boundaries of the groups (break points) defining classes. The Natural Breaks method may be suited for some applications in the natural sciences. However, Quantile classification is often a better starting point, because the method is easily understood and provides information about the shape of a distribution. The Quantile method breaks a distribution into classes—each with the same percentage of data points. For example, each quartile (quantiles with four classes) has 25 percent of the data observations, with the middle break point being the median.
By studying quantile break points, you can determine whether a distribution is roughly uniform (has equally spaced quantiles) or is skewed to the right (has intervals defined by break points that become progressively larger with larger values). The former become good candidates for the Defined Interval method (uniform distribution with easily read numbers for break points) and the latter for the Geometric Interval method (for an increasing-width intervals distribution of break points). Many attributes have skewed distributions.
Change the symbology method for the choropleth map from Quantile to Geometric Interval. The break points become larger to provide more detail for the long tail of the distribution. The break points for the quantile method were 830, 1620, 2970, 4831, and 11595. Those for the geometric method are 881, 2201, 4181, 7148, and 11595. Reselect the Gray (5 Classes) color scheme if necessary.
Try the Defined Interval method with an interval size of 2500. Although perhaps not the best method for this skewed data, this uniform distribution is easy to read, having “nice” numbers, multiples of 2500, with equal intervals.
Change the method back to five quantiles, close the Symbology pane, and save your project.
You will learn much more about 3D data and scenes in chapter 11, but you can easily convert a 2D choropleth map into a 3D scene to better visualize data. In particular, the map reader can get a better appreciation of extreme values relative to other values, which are not readily apparent by looking at only the color shading of choropleth maps. In 3D, features and layers are often physical features such as buildings, trees, topography, and so on, but you can display any numeric data as 3D. Here, you will learn how to convert a 2D map to a 3D scene and display neighborhood polygon features as 3D.
Your features and food stamp recipient data are now displayed in 3D.
Zoom by scrolling the mouse wheel, and pan by dragging the map. You will learn much more about 3D map navigation in chapter 11. Save your project.
Using ArcGIS Pro, you can display polygon data using point symbols in the center (centroid) of each polygon. In the next exercise, you will create a map showing the number of food pantries and soup kitchens in New York City neighborhoods as graduated size point symbols. The larger the symbol, the more food resources in each neighborhood. You can also display data as proportional symbols that are similar to graduated symbols but represent values as unclassified symbols whose size is based on a specific value.
The map of this exercise uses two copies of the Neighborhoods polygon layer. One copy displays household data as a choropleth map. The other layer displays the number of facilities using graduated point symbols. Using two copies of the same layer is a way to show two attributes of a polygon layer in the same map.
Notice that the interval width is uniform, 2, except for the last class, so this attribute may be a good choice for the Defined Interval method (uniform distribution). An interval width of 5 is a good choice to include the maximum value of 25. Equal-width intervals are the easiest to read and are, of course, best suited for uniform distributions, but this distribution is not essential. Another possible method is defined interval, which allows you to use easily read numbers such as 1, 2, or 5 times 10 to a power (for example, 0.1, 1.0, and 10).
Turn on and rename the second Neighborhoods layer as Under 18 receiving food stamps. Use Proportional Symbols, U18_Food as the Field, a shade of purple as the color, 2 as the Minimum size, and 20 as the Maximum size. Use the Bronx and Brooklyn bookmarks to study the relationship of food banks, soup kitchens, and persons over 60 and under 18 receiving food stamps in one map. Save your project.
A choropleth map showing population, such as the number of persons receiving food stamps, is useful for studying needs, such as the demand for goods and services. For example, delivery of food services for the poor requires capacities to match populations, including budgets, facilities, materials, and labor.
Choropleth maps of normalized population data have different uses than choropleth maps of populations. Dividing (normalizing) a segment of the population by the total population provides information about the makeup of areas. For example, areas with high proportions of total population receiving food stamps may be better candidates for food pantries and soup kitchens than those with low proportions, because the high-proportion areas are likely poor in many ways, including having poor geographic access to grocery stores and urgent health care.
In this tutorial, you will normalize the number of female-headed households (single mothers) with children under the age of 18 receiving food stamps by the total number of households in each neighborhood. You will find the same information for male-headed households (single fathers) with children under the age of 18 receiving food stamps and compare the two populations using a custom scale.
You will create a custom classification, which often is easier to read than other classifications. The Geometric Interval method works well for representing the long tails of distributions skewed to the right, but the break points of this method do not follow a pattern that can be read easily. The custom classification of this exercise has intervals that double in width (and therefore forms a geometric progression), which is read easily.
These settings show the fraction of single mothers with children under 18 receiving food stamps. Next, you will show the values as a percentage.
The numerical sequence of the custom break points has both increasing interval widths as desired for the long-tailed distribution and a recognized and easy set of values to read.
To easily compare two maps, especially normalized segments of the same total population, you will use the same numerical and color scales for both maps. ArcGIS Pro allows you to easily import symbology. Next, you’ll import and reuse the symbology of the female-headed households for the male-headed households. You will see many fewer male-headed households receiving food stamps.
Density maps divide populations and other variables by their polygon areas, yielding a measure of spatial concentration. If you divide a population by its polygon areas, the resulting population density (for example, persons per square mile) can provide information related to congestion or how people are distributed across an area.
A neighborhood with a high density of households receiving food stamps but a low density of food banks and soup kitchens may help determine potential locations for new food banks or kitchens.
Create a density map for “Persons below poverty level receiving food stamps (SQ MI)” using (POV_Food) normalized by Area_SQMI. Use Graduated Symbols; 4 classes, Manual Interval with upper values 300, 900, 2700, and 13000; circle 3 with Solar Yellow color; and a symbol size range from 4 to 12. Zoom and pan the map. Do you see any gaps with high population densities but low food facilities densities? Save your project.
You can place layers together into groups to manage them more easily. For example, you can change the visibility of an entire group layer with one click. You can also save a group layer (or any single layer) as a layer package (a file with an .lpkx extension). A layer package is one file with all data sources and symbology included. You can share this package with others, including in ArcGIS Online. In this tutorial, you will create group layers for populations and facilities in New York City by administrative and political features, including fire companies and police precincts. You will then create layer packages from layer groups.
There are two ways to create group layers. The first way is to create an empty group layer, and then add layers to the group layer. The second way is to select existing layers in the Contents pane and create a group layer of the selection. You will use the first method to create a group layer for police.
Here you use a shortcut to create a group layer using selected layers.
A layer package is a portable file with data and symbolization for all layers in a group or for a single layer that you can share or upload to ArcGIS Online. In addition to creating layer packages, you can also create and share individual layers as layer packages.
Create and add a layer package for NYC Fire to a new map, and save your project.
This chapter has four assignments to complete that you can download from this book’s resource web page, at esri.com/gist1arcgispro: