The True Colors of America’s Political Spectrum Are Gray and Green
2016 election results by precinct
Satellite image
2016 election results by precinct
With its lush green fields and trees, slate-gray roads and tiny blue backyard swimming pool, this aerial shot over Blue Ridge, Va., looks like any number of places in the United States.
It seems strange that an ordinary patch of land like this could offer clues about the political leanings of its inhabitants. But to some degree, it can. Think of it as the aerial-image version of those red-and-blue electoral maps.
Asked to describe this landscape, you might say it’s mostly green. Sorting the image’s pixels by color and brightness renders this impression even more precisely.
In the image (below at right) where this sorting has been done, the lightest colors from pavement and rooftops appear at the bottom. Moving up the picture, the grays give way to the deeper greens of trees and fields. At the top, the darkest bands of color represent the shadows in the landscape.
This photograph and its corresponding color palette resemble many other places in America, but over all, the United States is a patchwork of built and natural landscapes, with a variety of features and hues.
Farmland in the Finger Lakes region of New York State ...
… looks very different from the Mojave Desert in California.
The colors of this neighborhood in Boise, Idaho …
… are nothing like those of Lido Isle in Newport Beach, Calif.
Geographically distinct places sometimes share colors. The area around the Summerlin community in Las Vegas ...
… has a palette similar to this area near Coors Field in Denver.
Each aerial image above is a randomly selected snapshot of 65 acres’ worth of landscape in America. But what if these landscapes weren’t sampled so aimlessly? What might they reveal if derived from a purposeful sample — say, a political one, featuring areas with similar margins of victory in the 2016 presidential election?
This is a grid of aerial images taken across the contiguous United States, selected at random and arranged by political leaning. The neighborhoods on the left voted overwhelmingly for Hillary Clinton in 2016, while President Trump received outsize support in the landscapes on the right. Those in the middle were more evenly divided.
Aerial Images of Landscapes Across the Political Spectrum
+100 pts.
+80
+60
+40
+20
+20
+40
+60
+80
+100
Even
Margin of victory
Clinton won by more
Trump won by more
+100 pts.
+80
+60
+40
+20
Even
+20
+40
+60
+80
+100
Margin of victory
Clinton won by more
Trump won by more
+100 pts.
+80
+60
+40
+20
Even
+20
+40
+60
+80
+100
Margin of victory
Clinton won by more
Trump won by more
+100 pts.
+50
Even
+50
+100
Margin of victory
Clinton won by more
Trump won by more
+100 pts.
+50
Even
+50
+100
Clinton won
by more
Margin
of victory
Trump won
by more
You’ll notice a general trend in these images: from predominantly gray pavement at left to greener, more open spaces on the right.
But this grid represents less than one-tenth of 1 percent of the land cover of the country. To verify whether this gray-to-green, Democratic-to-Republican-leaning color trend applies to the contiguous United States, we had powerful computers process imagery of every square meter in precincts where votes were cast.
Below is what the result of that processing looks like. This image reveals the most frequently occurring 100,000 landscape colors in the United States, according to how people living in these landscapes voted in 2016.
The Colors of Clinton and Trump Precincts
+100 pts.
+80
+60
+40
+20
+20
+40
+60
+80
+100
Even
Margin of victory
Clinton won by more
Trump won by more
+100 pts.
+80
+60
+40
+20
Even
+20
+40
+60
+80
+100
Margin of victory
Clinton won by more
Trump won by more
+100 pts.
+80
+60
+40
+20
Even
+20
+40
+60
+80
+100
Margin of victory
Clinton won by more
Trump won by more
+100 pts.
+50
Even
+50
+100
Margin of victory
Clinton won by more
Trump won by more
+100 pts.
+50
Even
+50
+100
Clinton won
by more
Margin
of victory
Trump won
by more
The pattern we observe here is consistent with the urban-rural divide we’re accustomed to seeing on traditional maps of election results. What spans the divide — the suburbs represented by transition colors — can be crucial to winning elections. It’s part of why President Trump, seeking to appeal to swing voters, has portrayed the suburbs as under siege and menaced by crime. But the suburbs are neither politically nor geographically monolithic. They are where Democratic and Republican voters meet and overlap, in a variety of ways.
At each extreme of the political spectrum, the most Democratic areas tend to be heavily developed, while the most Republican areas are a more varied mix: not only suburbs, but farms and forests, as well as lands dominated by rock, sand or clay.
Landscapes Across the Political Spectrum
Water
Other
Farms
75% of land is...
Forests
50%
Developed land in
cities and towns
25%
+100 pts.
+80
+60
+40
+20
Even
+20
+40
+60
+80
+100
Margin of victory
Clinton won by more
Trump won by more
Water
Other
Farms
75% of land is...
Forests
50%
Developed land in
cities and towns
25%
+100 pts.
+50
Even
+50
+100
Margin of victory
Clinton won by more
Trump won by more
Water
Other
75% of land is...
Farms
50%
Developed land in
cities and towns
Forests
25%
+100 pts.
+50
Even
+50
+100
Clinton won
by more
Margin
of victory
Trump won
by more
Water
Other
Farms
75% of land is...
Forests
50%
Developed land in
cities and towns
25%
+100 pts.
+80
+60
+40
+20
Even
+20
+40
+60
+80
+100
Margin of victory
Clinton won by more
Trump won by more
The earlier Clinton and Trump color gradient takes the country into account. If we break down every state the same way, individual color palettes for the contiguous 48 states emerge, revealing local characteristics of each. Some have more Democratic-leaning gray areas, like New York, while others feature a relatively even distribution of colors across the political landscape, like Mississippi.
The images also reveal areas with lopsided electorates, like North Dakota and the District of Columbia. White areas in those images indicate a lack of landscapes belonging to that part of the political spectrum.
The Color of Clinton and Trump Precincts in Each State
Alabama
Arizona
Arkansas
California
Colorado
Connecticut
Delaware
District of Columbia
Florida
Georgia
Idaho
Illinois
Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Minnesota
Mississippi
Missouri
Montana
Nebraska
Nevada
New Hampshire
New Jersey
New Mexico
New York
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolina
South Dakota
Tennessee
Texas
Utah
Vermont
Virginia
Washington
West Virginia
Wisconsin
Wyoming
A closer look at Texas shows landscapes ranging from dense cities to rolling grasslands, forests and remote regions where oil and gas development dominate the terrain.
The Texas Landscape
Part of Lubbock, Tex.
Clinton +81 pts.
Chase Oaks, Tex.
Even
Jacksonville, Tex.
Trump +50 pts.
Clinton +100 pts.
Even
Trump +100 pts.
Part of Lubbock,
Tex.
Clinton +81 pts.
Chase Oaks,
Tex.
Even
Jacksonville,
Tex.
Trump +50 pts.
Clinton +100 pts.
Even
Trump +100 pts.
Part of Lubbock,
Tex.
Clinton +81 pts.
Chase Oaks,
Tex.
Even
Jacksonville,
Tex.
Trump +50 pts.
Clinton +100 pts.
Even
Trump +100 pts.
Massachusetts is a mostly liberal state; there were no precincts where Mr. Trump won by more than 30 percentage points. But even here, the trend of less pavement and increased green space in the more Republican-leaning areas is apparent.
The Massachusetts Landscape
Part of
Cambridge, Mass.
Clinton +83 pts.
Holyoke, Mass.
Clinton +47 pts.
Plymouth, Mass.
Even
Clinton +100 pts.
Even
Trump +30 pts.
Part of Cambridge,
Mass.
Clinton +83 pts.
Holyoke,
Mass.
Clinton +47 pts.
Plymouth,
Mass.
Even
Clinton +100 pts.
Even
Trump +30 pts.
Plymouth,
Mass.
Even
Part of Cambridge,
Mass.
Clinton +83 pts.
Holyoke,
Mass.
Clinton +47 pts.
Clinton
+100 pts.
Even
Trump
+30 pts.
No image of a single neighborhood or town can perfectly summarize a national political landscape. Here are 100 randomly sampled images of precincts across the country where Mr. Trump and Mrs. Clinton had equal support in 2016. There are some visual similarities, like the density of development, but there are also many differences, including the layouts of neighborhoods and the patterns of green space.
Places Where Clinton and Trump Split the Vote
That said, lookalikes do exist across the political spectrum. When some landscapes are compared from above, their resemblances are striking, but on the ground, the political leanings of the precincts they are part of couldn’t be more different.
Suburban Lookalikes
Washington, D.C. Clinton +83 pts.
Cranston, R.I. Trump +3 pts.
Patton, Ill. Trump +40 pts.
Rural Lookalikes
Lawton, Okla. Clinton +56 pts.
Hillsboro, Ore. Even
Taylor, N.Y. Trump +49 pts.
Lookalikes are anomalies, though, in the sea of palettes that we can now use to help us visualize the urban-rural voter divide. Thanks to our growing ability to process enormous amounts of data, that phenomenon is now more accurately expressed not in blues and reds, but in grays and greens.
When you move around the place you live in, think about what colors you see. Those hues may say something about how your neighbors (and even you) might vote this November.
And that patch of green we began with in Blue Ridge, Va.? The precinct it is part of went for Donald J. Trump by a margin of 51 points in 2016.
About this project
All images were derived from the National Agricultural Imagery Program (NAIP) and accessed via the Descartes Labs geospatial data platform. The Department of Agriculture collects high-resolution aerial imagery across the contiguous United States every two years. The images shown here were collected in 2016 and 2017. All imagery collection does not occur simultaneously; the process is split into multiple collections across states, which vary seasonally. Colors from multiple seasons may be incorporated into a single chart.
To generate each color gradient map:
• NAIP imagery was indexed to 2016 precinct-level election results, creating a searchable set of image tiles, half a kilometer across, that cover nearly every square meter of the contiguous 48 states.
• Images were sorted into bins based on 2016 voter margin. For each bin, each image tile was processed to create a histogram that shows how many times each unique color value appears.
• The most frequent 100,000 colors for each voter margin class were sorted by luminance, from least to most bright.
Tim Wallace is a Senior Editor for Geography at The New York Times.
Krishna Karra is a Machine Learning Scientist at WattTime.
Additional production by Alicia Parlapiano.