Data Visualization: Western U.S. Demographics

Home ยป Data Visualization: Western U.S. Demographics

Changing Demographics
of the West

ruralwest.stanford.edu/demographics

Essay

The Contraction and Expansion of the Rural West

Historical Background

What is Rural?

Data Visualization

Western U.S. Demographics

 

 


This visualization enables users to expore demographic changes over time and space using U.S. Census data at the county level from 1850 to 2008. The visualization inclides an overlay --- under the "urban overlay" tab -- to black out current urban and suburban areas of the country to see more clearly changes in rural areas.

A variety of demographic parameters can be viewed using the "Now Displaying" menu, and the "Display Mode" can be changed from a single year to two alternating years that can be chosen on the timeline. Hiding county boundaries can help users see larger patterns.

An Animated View of Demographic Change from 1850 to 2008

The visualization uses historical data and current estimates of the U.S. Census Bureau to map and animate demographic change in the United States.

  • The default map view is the Western portion of the United States, but the data cover the entire United States, and the map can be zoomed out or pushed around with the mouse to display any part of the country;
  • Each demographic parameter in the map appears under the heading "Now Displaying";
  • The “Display Mode” button allows for an animated comparison between two data points within the same parameters; for example, population density can be made to toggle back and forth between 1910 and 2008 by clicking "2 yrs alternating" and then dragging the timeline to those two dates;
  • “Display Map Overlay” highlights demographic change in rural areas by marking our urban areas on the map;
  • The play button in the lower left hand corner of the visualization animates the data through time from any chosen starting point.
  • The data are mapped by county, and the map can also be zoomed to that level.
  • County names, size, and current parameter values appear in the black box above the timeline when the mouse scrolls over a county boundary.

Visualization design, data mining, and programming by Dan Chang and Yuankai Ge. Concept and creative direction by Michael de Alessi. Creative consultation by Geoff McGhee, Krissy Clark, and Alex Braman. Editing by Jon Christensen. Citation for this visualization: Daniel Chang, Yuankai Ge, and Michael De Alessi, "The Contraction and Expansion of the Rural West: 1850-2008," from Visualizing the Rural West, April 2010, Bill Lane Center for the American West, Stanford University, http://www.stanford.edu/group/ruralwest/cgi-bin/drupal/content/rural-wes...

The Parameters

Explanations of the parameters with examples from the visualization are listed below. Click the parameter to see the example given in the visualization.

  • Population: Absolute population numbers by county.
    Example: Hitting the play button shows the westward expansion of the latter half of the nineteenth century, which then fragments and eventually concentrates the West Coast and parts of the Southwest.
  • Population density: The average number of people per square mile within a given county
    Example: Hitting the play button reveals a similar transformation as above, but density highlights the low population throughout the non-coastal West, apart from the Salt Lake City and Denver areas.
  • Percent of total population: County population as a percent of the total U.S. population.
    Example: Shows the increasing emptiness of the middle in relation to the rest of the country, and further highlights the concentration of people on the West Coast and Southwest.
  • County boundaries: The political-geographical boundaries of any given county.
    Example: Shows the proliferation of political boundaries, and therefore of political institutions, throughout the West between 1870 and 1920. By comparison, the change from 1920 to 2000 is negligible. Also, unchecking the box labeled "Show County Boundaries" hides the county boundaries in all of the maps and makes the animation more seamless.
  • Change in population since last census: The percent change in total population from one Census to the next
    Example: For example, comparing 1930 to 1940 captures the exodus from the middle of the country following the Dust Bowl.
  • "Frontier" counties: The "frontier" was considered the boundary between "unpopulated" counties with fewer than people per square mile and "transfrontier" counties with 2 to 6 people per square mile.
    Example: This was the definition of the "frontier" last used by the Census in 1880 and famously dropped because of the fragmentation of the frontier in 1890. As the visualization shows, the frontier was no longer demarcated by a line after 1880, but "unpopulated" and "transfrontier" areas have persisted, shifted, and morphed in the West.
  • Reverse population color palette: Areas of greatest decline in population are represented by the darkest colors.
    Example: This parameter emphasizes the empty places, and their persistence in the rural West and the Great Plains. It also points to the growth of the Southwest in particular.
  • Reverse population density palette : Areas of greatest decline in population density are represented by the darkest colors.
    Example: Comparing the same years as above (1910 and 2008) emphasizes the emptiness of the rural West, while the changes in the Great Plains appear less dramatic.
  • Population ages 0 - 19: Total population of people between the ages of 0 and 19.
    Example: This data begins in 1920, and shows the amazing demographic progression toward an aging population.
  • Population ages 20 - 44: Total population of people between the ages of 20 and 44.
    Example: This age range, which has been called the prime of productive life, becomes more evenly distributed over time in the East, but not so in the West.
  • Population ages 65 and up: Total population of people older than 64.
    Example: Again, the aging population begins to show up in the 1920s, and the map illustrates the remarkable aging of the population in the Great Plains.

 

Census Data

Map Data/Shape Files

Other Sources

Notes on Data and Visualization

The visualization “An Animated View of Demographic Change from 1850 to 2008” was created by graduate research assistants Yuankai Ge and Daniel Chang using Flex3/ActionScript[1], complemented by libraries of Flare[2] and ShapeFile Renderer[3] . The historical county boundary shape files, in Albers Equal Area projection, came from the Louisiana State University Historical US County Bounday Files (HUSCO).[4]

Historic data on population came from the Inter-University Consortium for Political and Social Research (ICPSR)[5] at the University of Michigan, and projected estimates for 2008 came from the U.S. Census Bureau[6]. These data were integrated into the HUSCO country boundary files using historical FIPS code.[7]

A word about county boundary data:

Bringing historic county boundary data into a modern, digital map context is a tricky business. Boundaries frequently shift as markers change over time, and political boundaries are often redrawn. The HUSCO boundary files are not the most accurate, especially at a very fine scale. For the Rural West project, however, we are interested in broad-scale changes, and so we decided that the loss in accuracy at finer scales was outweighed by the lower memory requirements of the HUSCO data (in other words, our map will load and play faster).

The most sophisticated data set on county boundaries (both chronologically and geographically) is supplied by the Newberry Library’s “Atlas of Historical Country Boundaries,” which identifies boundary changes to the day and geolocates them precisely.[8] These files, however, are not quite complete (Georgia is the only state remaining), and for the purposes of integrating decadal Census data, the exact date of a boundary change is not necessary. The next most geographically accurate county boundary shape files come from the National Historical Geographic Information System (NHGIS) at the University of Minnesota. Even the NHGIS files, however, are 5-7 megabytes for each year, as compared to 1-2 megabytes for the HUSCO data. Since our map covers 17 decades, the difference in upload time is significant. The compression tool Mapshaper was used to reduce the NHGIS shapefile size, but even at 40%, the distortion was greater than the HUSCO files.

Currently, the shapefiles and census data files (in .csv format) require uploading 30MB, which normally takes less than 10-20s before starting the visualization. Because we expect users of our maps to visualize broad trends, not to find the GPS coordinates of their nearest county boundary, we felt this tradeoff was acceptable.

1 http://www.adobe.com/products/flex/ http://www.adobe.com/products/flex/ and http://www.actionscript.org/,

2 http://flare.prefuse.org/.

3 M.W. Cobb, “The Jovial Farmer Boy (songbook)” (Chicago: The John Church Company, 1885). Accessed through the Library of Congress Performing Arts Encyclopedia, http://www.cartogrammar.com/blog/simple-shapefile-drawing-in-actionscript-3/.

4 C. Earl, S. Otterstrom, and J. Heppen, “HUSCO 1790–1999: historical United States county boundary files.” Baton Rouge, LA: Geoscience Publications, Department of Geography and Anthropology, Louisiana State University (1999). Online at http://www.ga.lsu.edu/husco.html.

1 M.R. Haines and ICPSR (The Inter-university Consortium for Political and Social Research). “Historical, demographic, economic, and social data: the United States, 1790–2000.” Ann Arbor, MI: Inter-university Consortium for Political and Social Research (2004). Online at http://www.icpsr.umich.edu/.

6 U.S. Census Bureau, “Population Estimates”. Online at http://www.census.gov/popest/archives/2000s/.

7 See “What is a FIPS code” at the Social Explorer website: http://www.socialexplorer.com/pub/help/FAQ.aspx#WhatIsFIPS.

8For background on the Newberry project and methodology, see John H. Long, “Atlas of Historical County Boundaries," The Journal of American History (March, 1995), 81(4): 1859-1863.

 

Last modified Wed, 8 Feb, 2012 at 15:15