Boundaries or Centroids?
This story map presents guidance to help you make the best choice between using boundaries or centroids (or both!) given your needs.
Over 100 feature layers containing the most recent five-year demographic and housing estimates of the American Community Survey (ACS) are available within the ArcGIS platform through ArcGIS Living Atlas of the World. These layers will be refreshed annually so they will always contain the most current ACS data!
When browsing through this group of layers , you'll see two options for each topic: boundaries and centroids. You may be asking, "what's the difference?"
Boundaries or Centroids? How to Choose?
The boundaries layers contain polygon data that show the administrative subdivisions. The centroids layers contain point data, so they are more geographically simplistic. With much less spatial data, the centroids will be slightly more performant in that they usually draw a split-second faster. Sometimes you will want to work with boundaries so that the administrative borders are available, and sometimes you will want to work with centroids to take advantage of the greater options that centroids allow.
Boundaries
Filled-Polygon Maps Require Boundaries
Boundaries allow for maps that show a filled polygon, so that the official borders of the state, county, or tract are visible.
For example, this map of Predominant Child Living Arrangements shows the tracts as filled polygons. Click on the tracts to see the breakdowns. Zoom out to see counties.
Boundaries
Graduated and Proportional Symbols are Still Available with Boundaries
What if you're working with a boundaries layer and you want to show two attributes using both color and size? The map will render as symbols, but the boundaries will be visible, especially when the map reader clicks on the pop-ups.
For example, in this map of Teen Employment displays counts with size and percentages with color, but the outline of the boundary can be seen when clicking anywhere within the boundaries of the feature, not just on the symbol. Click on a county to give it a try.
Centroids
Centroids Allow for More Symbology Options
From fun icons in the Places of Interest symbols to Firefly symbols to custom images and even gifs , point layers allow for a wide array of symbology options.
This map shows the count of naturalized citizens using an American flag as the proportional symbol, which comes from the People and Places icons.
Clustering Requires Centroids
Using clustering within ArcGIS Online can be a great way to explore your data, just keep in mind that clustering only works with point layers. The clustering option will adjust the clusters displayed in the map as the map reader zooms in or out. You can even adjust the pop-up to display the most predominant category in the cluster.
Analyses Summarizing Data is Easier with Centroids
Say you want to summarize the data for tracts whose centroids fall into a buffer or a drive time polygon from a school, hospital, or store location. Working with centroids from the start eliminates the need to first convert boundaries to centroids.
Or Use Centroids and Boundaries Together!
There are some great maps in the ArcGIS Online Group of Examples of Maps and Apps that use the American Community Survey feature layers in the Living Atlas that have combined both boundaries and centroids in creative ways.
This map of Predominant Highest Level of Education uses firefly symbology with the centroids layer with the counterpart boundaries layer underneath symbolized with the same color palette. The pop-ups are only turned on within the boundaries layer to avoid duplication. The high transparency setting on the boundaries layer and the dark basemap give this map a nice effect!
Predominant Highest Level of Education in the US (ACS)
This map of Predominant Commute Alternatives combines icons used as proportional symbols with boundaries underneath for a map that needs no legend!
What are the most popular commute alternatives?
For more information and ideas about using boundaries and centroids together in the same map, we have a blog post for you: