Company Blog

Posts Tagged ‘zach wilson’

Map of Swine Flu Cases — CDC Reported, 10:30am on 4/30/09

April 30th, 2009 by Zach Wilson

Hi.  We’ll update this post as new numbers become available.

Our data is from the CDC : http://www.cdc.gov/swineflu/

Please let us know if you see new numbers before we do.  Thanks!

note: colors by log scale

Map backgrounds are provided in both black and white in case you want to post them anywhere else. If you want to adjust colors, scale, view angle or other details, please use our software, which can be downloaded in a free version from our main page.

Please note, colors are by logarithmic scale.

Please note, colors are by logarithmic scale.

Where Obama Won — ‘08 Election County Analysis

November 13th, 2008 by Zach Wilson

Note: We gathered this data from major media outlets as they posted it, and it is not necessarily reflective of the most current tabulation. For the most recent and official data we suggest Dave Leip’s Election Atlas. Any of his data can easily be imported into our software.

http://www.uselectionatlas.com

Also, please note that what follows is a somewhat detailed handling of patterns in election data. For a more general overview you may want to scroll down to the next couple of posts. If you are going to skip down the page, you may want to quickly look at the last four images of this post anyway because Obama’s victory in Indiana, shown there, is a real standout among the many images that follow.

A prominent aspect of US election geography is that most urban areas favor Democrats and most rural areas favor Republicans. This past November 4th, while more overall counties voted Republican, Obama won because most of the most populated counties voted for him. We can see this pattern in the two images below. Flat counties in dark red are McCain victories, while all counties rising from the map are Obama victories, and the tallest and most blue are the places where Obama won the greatest percentage of the vote. With the purely color-coded map it is obvious how many more counties preferred McCain — all of the dark red ones. In the second image, it is much easier to see which areas favored Obama, and to what extent.

Arguably, a more interesting detail of the recent election is how, or rather where, Obama won that Kerry didn’t. We can begin to see where Obama gained votes by looking at voting trends from 2004 to 2008 in contrast with voting trends from 2000 to 2004. The images below use color to show the rate of change in the percentage of the vote going to the Democrat candidate. During each time period, blue counties have an increasing percentage of the vote leaning to Democrats, and in red counties the Democrat percentage is decreasing (which is to say red indicates an increase in Republican preference).

The image below shows that from 2000 to 2004 much of the Plains states, the South, and Utah, and Texas tended more Republican. In this case, blue counties are those that voted in a greater percentage for Kerry than they did for Gore.

Most of the next image down is blue, showing that most counties favored Obama against McCain to a greater extent than they had favored Kerry against Bush. The red counties preferred McCain more than they preferred Bush relative to their respective rivals. A band of counties stands out running from the Appalachians into Arkansas, also part of the panhandle of Florida, and western Louisiana. In addition we can see McCain’s home-state advantage — Arizona leaned slightly Republican this election in contrast with neighboring states.

We can extend the analysis further by looking more closely at a few regions and by adding another variable to the analysis. In the following images, colors will follow the same pattern as in the last two images, with red indicating an increase in the percentage vote for Republicans and blue as an increase in the percentage vote for Democrats. In addition height will show whether or not a county voted more than %50 for the running Democrat. All flat counties were Republican victories and all counties above the map were Democrat victories.

Specifically, in both of the next two images, color indicates the direction of change from 2004 to 2008, but in the first image, the counties above the map are counties Kerry won, and in the second image the counties above the map are counties Obama won. These settings mean that anywhere above the map in the first image and flat on the second image is a county that Obama lost. Obama’s relative losses are always red. By contrast, anywhere in blue, and above the map in the second image, but flat in the first image, is a county Obama added to the Democrat coalition.

Maybe you can see that Obama lost counties in parts of Florida and West Virginia and Kentucky and gained in North Carolina, Indiana, and other parts of Florida.  Don’t strain your eyes too much though, as we’ll take a closer look right away; first the nationwide images…

Looking more closely at the West we can see that Obama did broadly better than Kerry (shown by blue) whether or not the county was in aggregate a Democrat victory, and as noted earlier, McCain did slightly better than Bush in Arizona.

Moving east and focusing on the areas of Texas and West Louisiana where the Democrat share of the vote declined (in red), we can see that very few of these counties were won by Kerry because these areas are flat. By contrast some counties in Arkansas were won by Kerry (shown as elevation) but appear in red, which suggests Obama may have lost some of these counties that Kerry had won.

Looking from the west at Arkansas and into the Appalachians, we can see through a comparison of the next two images that some red counties switched from a majority favoring Kerry to a majority favoring McCain. In the first image we have counties above the map if Kerry won these counties. In the second image the counties above the map are counties Obama won. As a result, we know that Obama lost some counties that Kerry had won in this red band, but at the same time much of this band was Republican leaning to begin with and only became more so.

Looking from the south, we can see that Obama also lost a few counties that Kerry had won in Florida. In the second image, Jefferson Davis County stands out as an example of where Obama retained a majority of the vote despite winning a lesser share than Kerry. We can also see in this region that Obama increased the percentage vote won in places that already voted heavily Democrat.

We’ll now turn to examine a few states where Obama won new counties and earned these state’s electoral votes as a result. For example, this was the case in Florida. Obama lost counties in the panhandle but made up ground by winning other counties Kerry had not, a few of them being panhandle counties neighboring those he lost. The colors in the first image show the trend in voting from 2000 to 2004. In the second and third images, color indicates the trend direction from 2004 to 2008, and height shows, where Kerry won, and then where Obama won. Given the high population in Miami-Dade County (in the far south east and poping off the map in the final image), we can be sure it was important for Obama’s victory in the state, though the general blue trend of most counties is also notable.

A quick look at trends in Ohio shows us that Kerry and Obama gained ground in rather different counties, split along an east-west divide. The first image is the trend for 2000 to 2004. The second image is the trend from 2004 to 2008.

In North Carolina, we can see the trends 2000 to 2004 and then 2004 to 2008, just as above, except in North Carolina Obama’s gains were more consistent, improving the Democrat margin in most counties, matched by Kerry’s fairly consistent losses shown in the first image.

These widespread gains tipped the balance to gain the state for Obama. Below, colors remain as in the most recent image, showing where Obama gained ground, and height shows first where Kerry earned a majority, then where Obama was able to nudge above %50 of the vote. You may have to look closely to count them because mainly the differences are small, but every county that nudges off the map in the second image is a place that helped secure North Carolina for Obama.

Last, but possibly most striking, we’ll look at Indiana. These four images follow the same pattern as the last four. First, the trend 2000 to 2004. Second, the trend 2004 to 2008. Third we use height to show those counties won by Kerry. Fourth we use height to show those counties won by Obama.

Poverty and Politics - KML and UUorld

November 6th, 2008 by Zach Wilson

Our engine allows us to export KML files, several of which can be loaded simultaneously into Google Earth. In the following example, I created and imported two region lists: one for the counties won by Bush, and one for counties won by Kerry, in the 2004 presidential election. I loaded data on poverty rates, with the variables scaled as percentiles, and removed all but the top half of counties, leaving only those counties over the 50th percentile.

I made two new color spectra: one running from bright red to pale red, and the other from bright blue to pale blue. Finally, I used the blue colors for Kerry voters and red colors for Bush voters, such that counties in brighter colors and with taller extrusions represent the poorest counties that voted for the given candidate; those that are paler and shorter are less poor counties (although still poorer than average).

I exported the two studies as KML files. Here are the results…

First, the images in UUorld before export to KML:

Next, both KML files loaded into Google Earth…

Obviously in either mapping engine it is easy to zoom in on different regions for a closer look. In this case, using Google Earth, it is striking how the most poor counties in the US seem to vote Republican or Democrat in clusters. For example, around Jacksonville, Florida and in southwestern Kentucky some of the poorest counties in the nation favored Bush, while just west of Jacksonville and in central to eastern Kentucky Kerry was favored. See below.

And for reference, here is the western half of the country…

This last image is shot from the Northeast and I am compelled to wonder if it somehow captures the perspective of New Englanders when they think of poverty across the rest of the country.

We will be making similar images of the results from the recent election just as soon as we process the data. Also, for more images showing an experimental analysis on poverty and politics in the 2000 and 2004 elections, take a look here:

Elections and Poverty: Visual Analysis.


Seeing Sprawl through Housing Unit Ages

November 6th, 2008 by Zach Wilson

In the following images we can see that the centers of cities are older, shown as greener; and the suburbs are newer, shown in red. This makes sense, but it is one more case where a picture can help anchor what is already known. In addition, while we may know by intuition that suburbs have newer homes than the centers of cities, we may not know which suburbs are the newest unless we consult a map. Below we will see a number of patterns based on average housing unit ages by zip code.

San Francisco, and the Bay Area:

Washington D.C. Metro Area:

Chicago Metro Area:

In the next two images, we can compare two “twin cities” and hazard a guess about which of the two dyads are more fluidly connected. In the first image, it appears that Dallas and Fort Worth have two separate cores; whereas Minneapolis and St. Paul appear as a single oblong city.

Dallas and Fort Worth:

Minneapolis and St. Paul:

Looking at New York City below, we can see, as we would expect, that most of the city, and even the surrounding metro area, is rather old in comparison with only a few outlying patches, mainly on Long Island, that have seen more recent development…

Looking more closely at Manhattan, we can see that on average the Upper West Side has older houses, darker green, than the Upper East Side…

How old is your neighborhood?

Where Afghanistan Lives

October 4th, 2008 by Zach Wilson

Initially our software was designed for international comparisons and we only supported country-level data. It was a major aim of our current release to support more-fine-grained data, yet when we built the new capabilities we anchored exclusively on US data, easily available through the Census. Before release, I wanted to see if we could handle other sub-national boundaries, which led me to Afghanistan. Finding population data by province was not too hard; parsing it into an appropriate format took a bit more time, which explains the word “clean” in the image titles below — you are seeing my file name notation.

When I was finally able to load the data I was hungry to see the fruit of my labor.

And when this first image loaded up, I confess I was disappointed.

urbanity-afg-clean_11.jpg

Honestly, I had been hoping for more activity; I had not expected the flatness of all provinces outside of major cities. Then I realized, the very thing I was unhappy about was new information about Afghanistan, new for me at least: Kabul is a large city; so is Ghazi; Kandahar, in orange, is substantially smaller but still a major population center.

In addition, from exploring other data I knew that if the highest value(s) are clipped, one can see a layer below the initial image. In some cases clipping values aids in removing outliers deemed inaccurate. While that step should always be taken cautiously, remembered, and ideally noted, in this case I just wanted to see Afghan population aside from the major cities, the image of which I had already absorbed.

After removing Kabul, new texture showed up in the image:

urbanity-afg-clean_9.jpg

Next I clipped Ghazni, a city I had never heard of, although maybe I ought not admit such things. And, after removing Kandahar also, I could see even more demographic detail:

urbanity-afg-clean_2.jpg

Looking at the value labels for each province, I was surprised to see many with a population between 50,000 and 250,000 persons.

From the first image I thought that maybe Afghanistan was mainly barren, with all of the population clustered in the cities, but after drilling in farther I saw many population centers of middling size, spread almost evenly across the country, with some concentration in the north, and with only a small fraction of provinces having extremely sparse population.

On reflection, an image that seemed initially unappealing has taught me a few essentials about the demography of a place otherwise distant and foreign. Still I will wrestle with questions about which image is more accurate: the one that shows how far the the largest cities stand out, or the one which portrays the breadth of general settlement. As with many things, I suspect the answer is that both are valid perspectives; and that maybe even better is having both images at once, a task which maybe we can entrust to our memories.

Last, a word of encouragement to anyone experimenting with the software…

Please, Dive In!

We built the software to make it easy to explore and experiment with different ways of viewing whatever data interests you. And I would be remiss if I didn’t invite you to share what you find. Almost as new parents, we are eager for screen shots that show how our software handles any data that you decide to explore.