mapsbynik

Greater Austin Presidential Election Results 2016

This map depicts the results of the 2016 Presidential Election within the Greater Austin, Texas metropolitan area. It includes Travis, Williamson, Hays and Bastrop Counties, but not Caldwell*. Votes are shown as dots, where one dot represents 3 votes. For a more precise demonstration of voter preference, as well as their density and distribution, results are mapped at precinct level and dots have only been plotted within residential properties.

Votes for Hillary Clinton are colored cyan; votes for Donald Trump are colored magenta. Third-party votes are not shown. Brighter areas indicate larger concentrations of votes.

::

It's a pretty common political axiom in the United States to refer to cities as "islands of blue in a sea of red". The saying comes from the pattern produced by residents of urbanized areas tending to vote for Democrats (symbolized as blue since the 2000 election) along with rural voters tending to prefer Republicans (symbolized in red). On typical county-level post-election maps, these urban area stick out as small blue areas completely surrounded by red.

Geographers and demographers (and Democrats) tend to object to that depiction because it misleadingly implies that vast swaths of the country are filled with Republican voters. While it's true that rural counties tend to lean Republican (often heavily so), it's also true that those counties are often sparsely populated, consisting of vast tracts of uninhabited wilderness and farmland/ranchland. (See my Nobody Lives Hear map for an illustration of the unpopulated portions of the country). As the saying goes, "land doesn't vote".

Some cartographers try to compensate for this effect by distorting the representation of the land to reflect population counts, or using a gradient to indicate the winner's margin (or sometimes both). Others have tried making the map 3D, using height to show the total number of voters in each county. The result is a scattering of tall spikes showing urban voters and lots of flat low-population counties. Another intriguing, but underutilized, method is to fade counties in proportion to their population.

As an alternative approach for this map, I revisited the dot density style I began playing with earlier in the year. Though the map is drawn at the precinct level, the same principles apply as with county-level maps. Variations in precinct area and population can over- or under-emphasize the patterns of the local geography when using a less sophisticated presentation style. Indeed, precinct measurements illustrate the phenomenon perfectly. Precincts Trump won measure 2814 sq.mi. (7289 sq.km.) in area, while those that Clinton won measure a comparitively small 906 sq.mi. (2348 sq.km).

By drawing votes as dots and limiting them to the places people live, nuances in geographic distribution becomes plain to see. The "blue island" shines clearly, but "sea of red" is revealed to be the largely unpopulated space that it is.

This simplistic winner-by-precinct map shows the blue island/red sea concept. Despite being three times larger in area, the precincts won by Donald Trump contain significantly fewer voters than those won by Hillary Clinton.

Travis County, which has a reputation as a Democratic stronghold, shows that Hillary Clinton's strength was concentrated primarily in Austin's central urban corridor. She won many densely populated urban precincts with 80-90% of the vote. And while the residents in west Travis did lean toward Trump, they were fewer and more spread out compared to Clinton's voters. His highest portion of the vote topped out in the mid-60s.

North of Travis, Williamson County features an interesting gradient fading from blueish-purple in the southern portions of the county near Austin to the distinct magenta of the Trump-supporting suburb of Georgetown in the north, where he pulled about 70% of the vote.

Hays County was a virtual squeaker, with only 613 votes separating the two candidates. Clinton won respectable majorities in the developed cities along Interstate 35 in the southeast portion of the county (56+% in Kyle and Buda), and dominated San Marcos, home to Texas State University. There, she took 70-75% of the vote. But it wasn't quite enough to counter the county's rural areas where Trump won 60-65%.

Finally, in Bastrop County, Clinton won narrow majorities (low 50s) in the more developed areas, namely Bastrop proper, Elgin and near Elroy on the border with Travis County. The remainder of the county went heavily for Trump, scoring totals between 55-75%.

::

Further examination reveals some potentially promising trends for Texas Democrats. Hillary Clinton exceeded Barack Obama's 2012 vote percentages in Travis, Williamson and Hays Counties by 6.2, 3.5 and 3.1 points, respectively. She even beat his blowout 2008 margin in Travis by 2.8 points. Bastrop, however, was not as favorable to her, losing 3 points compared to Obama 2012. It's a small county though, and those lost votes were offset by increases in the other counties. Across the four-county area, Clinton showed a 4.2 point improvement over Obama's 2012 results and a 0.5 point increase over his 2008 numbers.

The picture is not so rosy on the Republican side. He may have won the State of Texas, but Trump's performance across these four counties was down significantly compared to Mitt Romney or John McCain. He beat McCain's numbers in Bastrop by 4 points, but that was his only improvement over previous elections. Romney beat his results in Bastrop by 0.3 points. In Williamson, Trump underperformed Romney by 8.3 points. In Hays, it was 6.6 points and in Travis, 8.8 points. Trump underperformed McCain by 4.6, 2.3 and 6.9 in Williamson, Hays and Travis, respectively. Overall, his percentage of the vote fell 5.1 points short of 2008 and 8.3 points short of 2012.

That 8.3 point drop in support seems to have been evenly split between Clinton (her +4.2 over Obama) and third-party candidates, who increased their share from 2.5% to 6.8% (+4.3 points) this year compared to 2012.

::

I found this to be a fascinating look at the Capital of Texas's electoral geography for 2016. I hope other folks are as intrigued by the visualization as I have been. In the metro area, there's a some red, a lot of blue, and a fair amount of purple.

*OMB includes Caldwell in its definition of the Austin-Round Rock Metropolitan Statistical Area. I don't have property zoning data for it though, so it has been omitted. Plus, with only 11,486 total presidential votes cast, those results would have barely been visible on the map.

Austin Proposition 1 Results Map

Proposition 1 was an industry-sponsored citizen's ballot initiative concerning recently-enacted city ordinances governing the operation of so-called ride hailing or transportation networking companies within Austin (think Lyft, Uber, and the like). A special referendum was held May 7, 2016 directly asking the voters in the city to overrule the city council and repeal the ordinances. A "Yes" (for) vote favored repeal, while a "No" (against) vote favored keeping the rules in place.

After an expensive and contentious campaign, the proposition was defeated 55%-45%. Voter turnout was 17.5%.

Examining the map, the most striking feature is the stark contrast between the heavily "Pro" areas around Downtown and UT-Austin and the heavily "Against" Central Austin neighborhoods. The map also shows some decent contrasts in South Austin, where some precincts voted strongly in favor while adjoining precincts were strongly opposed.

Further study is needed to explain some of these patterns.

Full-size map

It’s the 2016 primary election season in the United States and this map represents a small part of it. It depicts the geographic distribution of votes for Hillary Clinton and Bernie Sanders in Travis County, Texas (home to Austin, the state’s capital city) during the Democratic Presidential Party on “Super Tuesday”, March 2.

The dots on the map represent voters. They were generated from the precinct-level election returns published by the Travis County Clerk’s office. Yellow dots represent votes for Sanders, while turquoise dots represent votes for Clinton. To give a sense of where groups of supporters live, the dots are constrained to cadastral parcels zoned for residential use. Dots are drawn randomly within each precinct and parcels; they do not represent the locations of individual voters.

Despite losing the state, Sanders won the county by about 4600 vote. Not a great surprise, given the county’s left-leaning reputation. Some general observations:

Clinton actually won more precincts, 130 to 109.

But those wins were of smaller magnitude compared to Sanders’. Clinton’s precincts were won by an average of 78 votes while Sanders averaged nearly double the margin with 142 votes per precinct.

Geographically, Sanders showed extremely strong support in downtown neighborhoods, especially at UT–Austin and its surroundings. In the precinct housing most of the dorms, Sanders ran up a huge 5.5X, +1419 vote win. Impressive margins for Sanders were also found in the West University (+924), Pleasant Valley (+759) and Parker Lane (+622) neighborhoods.

Clinton scored the most lopsided victory, 52 to 8 (6.5X margin), in Webberville in the southeast portion of the county.

Clinton’s strongest areas of support came from the neighborhoods along RM 2222 west of Mopac, and on both sides of the river, scoring a handful of solid +200 to +300 vote victories. Additional strong showings were found at the Circle C Ranch developments in the southwest of the county and the Pecan Springs Springdale neighborhood in eastern Austin.

::

Election maps are always fun and I enjoyed putting this one together. The use of parcel data to limit the geographic extent of the symbology was a bit of an experiment for me. I think the results turned out well, revealing more detailed patterns among the candidate’s supporters than would have been revealed otherwise.

P.s. - For the curious: yes, I did look at the Republican Primary results as well. They weren’t that geographically interesting though, so I’m not likely to publish a map about it.

The Rally's – Checkers Divide (2014)

This map is a quickie to meet demand from the reddit thread about the Hardee's – Carl's jr map. It shows the distribution of Checkers and Rally's restaurants across the United States. Both chains were founded independently in the 1980s and merged operations in 1999.

No plans to do any analysis of the results, but at first glance, there does not appear to be a hard trend to the placement of the restaurants. Generally though, Checkers has some strong clusters in the South, but also in the NYC and Detroit areas. Rally's has a strong presence in the old Northwest Territory and in California, with a few islands scattered throughout Checkersland.

A few notes:

The closest Rally's-Checkers pairing occurs in Pine Bluff, Arkansas. The Checkers on UA-Pine Bluff is but 2 km (1.2 miles), a short 5 minute drive, to the nearest Rally's. There's another Rally's 4.3 km from campus.

Texas has one Checkers location in Beaumont, but residents of El Paso are actually closer to the Rally's locations in Arizona.

Chicagoland has both chains. Mostly Checkers in Cook County, with one Rally's in the south of the county. Lake County, Indiana has three additional Rally's locations.

One could travel over land from Sault Ste. Marie, MI to San Diego, CA and technially always be closer to a Rally's than to a Checkers.

See full-size map.

This map depicts the distribution of more than 2800 Hardee's and Carl's Jr restaurant locations in the United States (Alaska and Hawaii exluded). Hardee's locations are highlighted in red; Carl's Jr locations are shown in yellow. Areas shaded maroon are nearer to a Hardee's restaurant while the yellow-greenish color represents Carl's Jr territory. Neither franchise is found in states with grey counties.

Though owned by the same parent company, CKE Restaurants, the distribution patterns of the two chains is quite different. A visual review of the map reveals that the 1738 Hardee's locations in the dataset are spread more evenly across the franchise's territory, while Carl's Jr 1142 restaurants are generally clustered into a few high-density regions. With nearly 2800 locations in 42 states (sorry New England), the two chains combined make up the fourth largest fast food enterprise in America.

The region where the two brands come closest to each other is in northeast Oklahoma and northwest Arkansas (see the “Mixing Zone” map). At 30.7 miles (49.5 km) apart, the two closest restaurants of each chain are located in Siloam Springs, AR (Hardee’s) and Tahlequah, OK, inside the Cherokee Nation (Carl’s Jr). As sometimes the company airs the same commercial, but with a different name at the end, the advertising in this region must sometimes get confusing and must be the stuff of a brand manager's nightmare.

Despite being under the same umbrella for 17 years, the service areas of two chains remains distinct and separate. Neither chain has permanently encroached into the other's territory, creating a sharp east-west division. And I love the emphasis that this map places on that divide. The distribution of the two chains, as well as their corporate history, offer a near perfect illustration of how physical geography can guide economic activity and human settlement patterns.

In the west, Carl’s Jr restaurants are concentrated in urban and suburban areas. Ninety-five percent of the chain’s locations are found inside a Census Bureau designated urban area or urban cluster (defined by population density). And the typical Carl’s Jr location has 14,652 people living within a one-mile radius.

By comparison, only eighty-three percent of Hardee’s, a business founded to serve more rural populations, are located in urban areas or clusters. And those areas are generally smaller. At a one-mile radius, the typical Hardee’s restaurant is surrounded by only 4036 people.

These characteristics are the results of a half-century trajectory for each chain, where different opportunities were presented to them by the continental landscape. For Carl’s Jr, those opportunities were found in the boom times and population growth of the comparatively few habitable places in the American West. In Hardee’s case, opportunity was found among the smaller towns spread out among the eastern geography that supported their existence.

::

The eastern half of the United States is privileged to be situated with attributes advantageous to population growth and economic development. The combination of fertile soils, abundant precipitation, thousands of miles of navigable rivers, and relatively flat terrain on a continental scale provide tremendous support for widespread habitable living. As the nation grew, these physical traits allowed and encouraged the population to establish vast agricultural holdings, with small settlements peppered throughout as local hubs of economic activity. By the 1960s, these towns represented a large untapped market for Hardee's to target.

As the Hardee's franchise grew from its home in Rocky Mount, NC (population 32,147 in 1960), the chain deliberately set up shops in those smaller towns and rural areas. American car culture had been spreading east from its roots in Southern California during the previous decade. Drive-in fast food restaurants followed shortly, becoming common sights in the larger urban and growing suburban areas of the country. Less populated places, however, were excluded from the initial industry boom, and Hardee's saw the opportunity to bring the fast food experience to them.

Carl's Jr, on the other hand, got its start in Anaheim and Los Angeles, amid the post-war prosperity and growth that allowed the automobile’s dominance to alter the landscape. Massive engineering projects had brought water to the deserts of SoCal. At first, there were farmers, but as more people moved to the region, drawn by the tremendous boom times, those farms became suburbs. By the early 1950s, Anaheim was the fastest-growing city in one of the fastest-growing regions in the country. And pioneers like Carl’s Jr, McDonald’s and Taco Bell were there to kick start the industry.

But even the grandest irrigation schemes and engineering ingenuity can’t completely overcome nature. There’s only so much water to spread around before the beds run dry and only so much land to expand into before mountains start getting in the way.

Green areas on this map from USGS receive more precipitation than the landscape can use. Red areas demand more precipitation than is provided by the atmosphere.

Beyond the rough terrain, one of the principle factors affecting the west is the lack of water compared to the east. Evapotranspiration is the measurement of the amount of water that travels from the surface of the planet into the atmosphere. West of about 96.4 degrees longitude, evapotranspiration begins to outweigh precipitation for expansive stretches of the country. More water leaves the earth than is provided by the atmosphere, making agriculture more difficult than in the east. The arid conditions, combined with the lack of major rivers and the numerous deserts and mountain ranges, means that the types of settlement patterns seen in the east simply are not, and will likely never be, found in the west.

This assessment is backed up by statistics from the Census Bureau. As of the 2010 count, 13,186 Census-defined places east of 96.4 degrees had a population of less than 2500. Those 13,186 places were home to 10,298,685 people in total. West of that meridian however, 6046 places were home to only 3,657,049 people. Not only does the west have fewer small towns, those towns support smaller populations on average.

From this data, it’s clear that the Hardee's approach would not have worked as well in the west. Fewer small towns and smaller populations make them less attractive options for business expansion. In spite of millennia of indigenous habitation, 500+ years of European colonization, and more than a century of U.S. Homestead Acts (which essentially gave away land to settlers), the west is so generally hostile to survival that populations tend to be more densely clustered in the relatively few places where people can live. Thus we see that the growth of the Carl’s Jr franchise occurred in those clusters.

::

The difference in geography didn’t just affect the distribution patterns of the restaurants; it also played a factor in the merger of the two companies. Why did CKE choose to maintain two separate identities? Why hadn't Hardee's locations simply become Carl's Jr during the takeover?

That was plan when CKE bought Hardee’s in 1997. The company wanted to compete nationally with Wendy's, Burger King and McDonald's under the Carl's name. Carl's Jr existed in the west and Hardee's had locations in the east and the midwest. On the map, it looked like a perfect marriage, a fast way to expand from coast to coast. Had all Hardee's restaurants been converted, Carl's Jr would have indeed become the fourth-largest restaurant chain in the country.

Alas, these grand visions were not to be. Opposition to the changeover was strong. Hardee's executives, franchise owners and customers balked. The handful of eastern test locations eventually reverted to the Hardee's name. And of the few outright conversions were that were made, lasting success was limited to a few locales in the west (Oklahoma in particular took to the changes, possibly explaining the bulge in Carl’s territory around Tulsa).

By 2001, the parent company had settled on a hybrid approach (which continues in 2015). Elements of Carl’s Jr (such as certain burgers) were sent east and the long-time icon of Carl’s Jr—that happy little star—was also gifted to Hardee's as part of a rebranding campaign. From that time forward, the restaurants had two different names, but shared one logo. In the meantime, parts of Hardee’s menu (such as its breakfast items) went packing west.

The conversions didn’t stick, in part, because the culture surrounding Carl’s Jr didn’t mesh well with Hardee’s branding. By the nature of its rural, small town focus, Hardee’s had earned a following among an older, hard-working, blue-collar crowd with “folksy” values. Despite its troubles in the 90s, the Hardee’s brand still held value to its customers.

Carl’s Jr, as a product of Southern California, reflected a younger, more hip, more urban attitude. When CKE attempted to swap the down-home hangout for the unknown upscale joint with the awkward name, the company risked losing a dedicated customer base, one that had stuck with Hardee’s through the hard times.

It took a few years of growing pains to settle on a working model, but by many accounts, this match made on the map continues to succeed 18 years later.

::

I made this map to settle a curiosity I've had for more than a decade, from the time I first spotted a Carl's Jr location during a cross-country drive in 2002. As I travelled westward through Texas, I remember distinctly the restaurant’s logo catching my attention.

At the time, Hardee’s was in the midst of its revitalization and recovery following a disastrous turn in the 1990s. That star had been near the center of the recent campaign to turn the store around. Nearly everything in the South that said “Hardee’s” had the star next to it.

To my growing perplexity, more restaurants with more smiling stars greeted me in town after town, as I continued along Interstates 10 and 20. In my mind, the star belonged to Hardee’s (it had even been called the “Hardee’s Star” in recent ads). What was this other restaurant with the funky name doing with its logo?

It wasn't until much later (in those pre-Wikipedia days) that I learned the story of the purchase and merger of the two companies. When I found location data for the restaurants last year, I finally put to rest my questions about where one chain ended and the other began.

Full disclosure: I haven’t been to Hardee’s since the 1990s and have never stepped foot inside a Carl’s Jr.

Howdy folks! It's been a quiet few months around here, hasn't it? I assure you more maps are on the way. It is not a fast process, though. But rest assured, they are coming.

In the meantime, prints (yes, physical copies!) of the Nobody Lives Here map are now available. Printing is provided by Imagekind.com, which offers a plethora of sizing, canvas, and framing options. From the petite 18"x15" to the gigantic 68"x48", there's bound to be a format to fit a map aficionado's nook, cranny, home, dorm, office or billboard.

Nobody Lives Here prints are available in two options. The first is a vertical (portrait) poster of all 50 states drawn to equal-area scale, similar to the online version. It includes a explanation and discussion of the patterns present on the map.

The second version is a horizontal (landscape) depiction focused on the contiguous 48 states. It omits the discussion text (as well as Alaska and Hawaii) in favor of taking a more detailed view of the U.S. mainland.

Waffle House by Latitude

After the seriousness of the Nobody Lives Here map, we're back to the silliness of Waffle House. I promised more visualizations of WaHo geography, and here I deliver.

This graphic shows a mash-up of two visualization types: a map and a bar chart. The concept is similar to Bill Rankin's population histograms, which tabulate population counts by latitude or longitude. Here, Waffle House restaurant locations are counted by their latitude, grouped into segments of 0.25 degrees.

WaHo's status as primarily a southern institution is well-established. And the results here match that expectation, with only a handful of restaurants to be found north of the Mason-Dixon line. But these results also emphasize how Waffle House distribution is more a phenomenon of the Deep South. Of the 1606 locations in my (2012) data set, 66% of them are located south of the Georgia-Tennessee border (35° 59' N).

Looking at the longest bars, we see the Atlanta metro area dominating again, just as it did on the 3d Density Map. The first, third and fourth-longest bars pass straight through the ATL and its environs (Athens helps a bit too). There's another, smaller concentration near 33° N that we can attribute to locations in Dallas-Fort Worth. Nashville is also responsible for a relatively long bar.

The takeaway from this graphic for me though, is the bar that passes along the Gulf Coast (as well as the two smaller bars on either side). It's the second-longest bar on the page, yet no single city can account for it. Instead, it's mostly a string of locations tied together by one common element: Interstate 10.

The Gulf Coast between Jacksonville and Houston has 78 Waffle House locations within one mile of I-10 (and an additional 9 along the I-12 bypass of New Orleans). Plus, many of the towns along that route have restaurants near the coast. US 90 between Mobile and LaFayette adds an additional 19 locations.

Given the linear spread of Waffle House locations along the Gulf Coast, I'm tempted to coin the term "Waffle Belt" to describe that region. Actually, that doesn't sound too bad.

I hereby christen the Gulf Coast from Houston to Jacksonville as The Waffle Belt. May its syrup always be heated and its batter forever fluffy.

::

Map Notes

I chose the Miller Cylindrical projection for this map because I needed the parallels to be, well, parallel. Can't have curved latitudes with straight bars. I also wanted to minimize the area distortion in the northern states, so that ruled Mercator right out.

As with the 3D map, the Waffle House location data is my own creation made by geocoding addresses with Google and MapQuest.

Made with QGIS, Photoshop, Apple Numbers and my own Python script

Highway data from National Highway Planning Network, U.S. Dept. of Transportation

France Edition: Nobody Lives Here

This is not a "map by nik", but one by twitter user @matamix. And it is a fine companion piece to my United States version that I thought I'd share.

Based on population data from INSEE, it shows the parts of Metropolitan France where people don't live. Like the U.S. map, terrain plays a big role in determining where people settle. For instance, the roughness of Alps (southeast) and the Pyrenees (southwest) Mountains, prevents habitation. On the southern Atlantic coast, forestation acts as an impediment to human living in Aquitaine.

The vast expanse of "uninhabited" land in the northeast (Burgundy, Champagne-Ardenne) is a bit of a mystery. Much of it is arable land and likely covered in agriculture, especially on the side near Paris. But Brittany and Pays-de-la-Loire are also prime agricultural land. So why do we see such different settlement patterns between the regions?

Perhaps someone more familiar with the geography of France can clue me in. I'd appreciate it.

Update: jedo52 explained to me that the regions did not share agricultural practices historically. Brittany and the northwest are traditionally covered in a bocage landscape where housing is dispersed in many hamlets. In the east however, farm fields are more open with housing more concentrated into villages.

Finally, the numbers. France checks in at about 32% unoccupied compared to 47% in the U.S. Considering the higher population density in France, it's not surprising that the country is "less empty" of humanity.

Nobody lives here: The nearly 5 million Census Blocks with zero population

A Block is the smallest area unit used by the U.S. Census Bureau for tabulating statistics. As of the 2010 census, the United States consists of 11,078,300 Census Blocks. Of them, 4,871,270 blocks totaling 4.61 million square kilometers were reported to have no population living inside them. Despite having a population of more than 310 million people, 47 percent of the USA remains unoccupied.

Green shading indicates unoccupied Census Blocks. A single inhabitant is enough to omit a block from shading.

Update Jan 2015: Prints and canvas of the Nobody Live Here map are now available.

Update 2014.05.01: I've received a couple questions about Canada. Just to be clear, this map is of the United States only. It is based on 2010 data published by the U.S. Census Bureau, which for reasons I hope are apparent, does not include data on our friends in the Great White North. For a similar depiction of Canada, see this map whipped up by Michael Chung.

Map observations

The map tends to highlight two types of areas:

places where human habitation is physically restrictive or impossible, and

places where human habitation is prohibited by social or legal convention.

Water features such lakes, rivers, swamps and floodplains are revealed as places where it is hard for people to live. In addition, the mountains and deserts of the West, with their hostility to human survival, remain largely void of permanent population.

Of the places where settlement is prohibited, the most apparent are wilderness protection and recreational areas (such as national and state parks) and military bases. At the national and regional scales, these places appear as large green tracts surrounded by otherwise populated countryside.

At the local level, city and county parks emerge in contrast to their developed urban and suburban surroundings. At this scale, even major roads such as highways and interstates stretch like ribbons across the landscape.

Commercial and industrial areas are also likely to be green on this map. The local shopping mall, an office park, a warehouse district or a factory may have their own Census Blocks. But if people don't live there, they will be considered "uninhabited". So it should be noted that just because a block is unoccupied, that does not mean it is undeveloped.

Perhaps the two most notable anomalies on the map occur in Maine and the Dakotas. Northern Maine is conspicuously uninhabited. Despite being one of the earliest regions in North America to be settled by Europeans, the population there remains so low that large portions of the state's interior have yet to be politically organized.

In the Dakotas, the border between North and South appears to be unexpectedly stark. Geographic phenomena typically do not respect artificial human boundaries. Throughout the rest of the map, state lines are often difficult to distinguish. But in the Dakotas, northern South Dakota is quite distinct from southern North Dakota. This is especially surprising considering that the county-level population density on both sides of the border is about the same at less than 10 people per square mile.

Update: On a more detailed examination of those two states, I'm convinced the contrast here is due to differences in the sizes of the blocks. North Dakota's blocks are more consistently small (StDev of 3.3) while South Dakota's are more varied in area (StDev of 9.28). West of the Missouri River, South Dakota's blocks are substantially larger than those in ND, so a single inhabitant can appear to take up more space. Between the states, this provides a good lesson in how changing the size and shape of a geographic unit can alter perceptions of the landscape.

Finally, the differences between the eastern and western halves of the contiguous 48 states are particularly stark to me. In the east, with its larger population, unpopulated places are more likely to stand out on the map. In the west, the opposite is true. There, population centers stand out against the wilderness.

::

Ultimately, I made this map to show a different side of the United States. Human geographers spend so much time thinking about where people are. I thought I might bring some new insight by showing where they are not, adding contrast and context to the typical displays of the country's population geography.

I've all but scratched the surface of insight available from examining this map. There's a lot of data here. What trends and patterns do you see?

Errata

Due to a cartographic mishap, the Gulf of California was missing from the original version. Though it was quickly fixed, that version keeps popping up across the Internet. It displeases me I see it, yet I'm amused that people keep reposting it without noticing the error. The geography gods judge those people harshly.

Some islands may be missing on the hi-res edition if they were not a part of the waterbody data sets I used.

::

Waffle House Density by Quad

The map shown here is a truly quick-and-dirty visualization of the distribution and density of Waffle House locations as of summer 2012. It started as a way for me to play with 3D in Google Earth, but that giant spike in Atlanta was interesting enough that I thought it worth sharing.

The rectangles represent USGS 30x60 minute (1:100,000) map quad boundaries and the height extrusions are proportional to the number of Waffle House locations found within that quad.

Map Observations

Atlanta, Georgia is the clear standout on the map. Its quad contains 132 locations. Immediately north of Atlanta, the Cartersville, Georgia quad is next highest on the list with 45 locations. These are followed by Greenville, SC (34), Athens, GA (29) and Fort Worth, TX (25).

The Atlanta metro area's dominance on the map is not unexpected, though its scale is somewhat surprising. The first Waffle House restaurant opened in the Avondale Estates, an inner Atlanta suburb, in 1955 and the company is headquartered in Norcross, another nearby suburb.

Waffle House has a reputation for omnipresence in the south and that seems especially true in the Atlanta area. The Atlanta quad has nearly three times the number of restaurants as the next highest. And that next highest quad, Cartersville, actually contains most of the north Atlanta metro area. Atlanta is home to a full eleven percent of WaHo locations.

Beyond Atlanta, the map shows that Waffle House is indeed primarily a southern phenomenon with a few surprising and unusual patterns in a couple metro areas. Phoenix and Denver have a handful of locations, but there are none in California. Amarillo has five locations, but San Antonio has zero despite being a much larger city at the junction of two major interstates.

Top 21 Quads

Quad Name Number of Locations Atlanta (GA) 132 Cartersville (GA) 45 this quad includes the N. ATL suburbs Greenville (SC) 34 Athens (GA) 29 Fort Worth (TX) 25 Nashville (TN) 24 Griffin (GA) 23 Commerce (GA/SC) 20 Mobile (AL/MS) 20 Louisville (KY/IN) 18 Dallas (TX) 18 Macon (GA) 18 Bay Minette (AL/FL) 18 Charlotte (NC/SC) 17 Aiken (SC/GA) 17 Jackson (MS/LA) 17 Raleigh (NC) 16 Little Rock (AR) 16 Gulfport (MS/LA) 16 Biloxi (MS/AL) 16 Pensacola (FL/AL) 16

Background

This visualization stems from my irrational, lifelong infatuation with Waffle House, the always-open dive joint that manages to provide a low-cost, consistently-satisfying breakfast meal at its 1600+ locations. It's one product of a casual analysis I started a few years ago to find WaHo locations that were unexpectedly close to each other, and it snowballed into a bunch of different depictions of Waffle House geography. I recently stumbled on these old files and will (eventually) be sharing them as I get them cleaned up and presentable.

Additional Notes

Grouping by USGS 1:100,000 quads was an attractive option for this data for a couple of reasons:

The quads are continuous across the entire USA. There are no gaps, so every Waffle House location falls within one quad.

Quad boundaries are static. They will not have changed if I update the data.

Quads are named for the area they cover, making it easy to discuss particular locations.

At approximately 2392 square miles each, quads are a nice "bucket" size for tabulating this kind of location data meaningfully.

Paris Métro Coverage

There's a saying–common in travel books–that Paris is so friendly to pedestrians, no point in the city is more than 500 meters (about six minutes on foot) from a metro station. Putting that notion to the test, this map highlights areas where the saying is indeed true. Within the départment of Paris proper, 96% of its area (excluding the large parks) is but a short walk from any RER or Métro stop.