Purple America Solved

Description

Write a program to visualize U.S. presidential election results.

Historical context. During coverage of the 2000 presidential election, Tim Russert coined the political terms red states and blue states to refer to states that predominantly vote for the Republican presidential candidate (red) or the Democratic presidential candidate (blue). The news media use red-state blue-state maps, such as the one below, to display election results.

The problem. On this assignment, you will create more refined (and less polarizing) choropleth maps by mashing up two sources of data: geographic boundary data and election return data.

Geographic data. We supply geographic data (sourced from the U.S. Census) that describes the boundary of each state and county in the United States.

• The first line consists of four real numbers, representing the bounding box of the region. The first two numbers are the minimum longitude and latitude values; the second two numbers are the maximum longitude and latitude values.
• The next line is an integer that specifies the number of subregions.
• There is one block for each subregion (with a blank line separates blocks):
  • The first line of a block is an integer N that specifies the number of points in the polygon describing the subregion.
  • The second line of a block is a string that is the name of the subregion.
  • The third line of a block is a string that is the name of the region.
  • The remaining N lines of the block describe the polygonal boundary, given as N pairs of real numbers, representing the longitude and latitude coordinates.

We note that the number of subregions in USA.txt is not 50 for two reasons: first, we do not include either Alaska or Hawaii; second, we include an entry for each polygonal subregion—some states (such as Michigan, Florida, and California) comprise several polygonal subregions.

Election return data. We supply election return data (sourced from Dave Leip’s Atlas of U.S. Presidential Elections) that describes the results for each presidential election, by state and county. Each row consists of four fields, separated by commas: the name of a subregion, the number of votes for the Republican candidate, the number of votes for the Democratic candidate, and the number of votes for the Independent (or third party) candidate.

You can download all of the geometric and election return data files collectively as purple-america-data.zip.

Part 1. Write a program White.java that takes the name of a region as the command-line argument and produces an outline map, as in the examples below:

% java White NJ          

% java White USA-county

For simplicity, draw the point with longitude x and latitude y at location (x, y) in the plane. Use the bounding box of the region to determine the part of the plane to display in the drawing window and rescale the coordinates accordingly.

Part 2. Write a program RedBlue.java that takes two command-line arguments (the name of the region and the year of the election) and produces a red-state blue-state map, as in the examples below:

% java RedBlue NJ 2012   

% java RedBlue USA-county 2012

Part 3. A more refined visualization reveals that the United States is not as polarized by geography as suggested in the above visualizations. In 2000, Bob Vanderbei created a Purple America map, in which each region is colored in a shade of red, green, and blue, according to the proportion of votes received by each candidate. Specifically, if the Republican, Independent, and Democratic candidates receive a₁, a₂, and a₃ votes, respectively, then we draw the subregion using the following formula:

% java Purple NJ         

% java Purple USA-county

Extra credit. There are limitless opportunities for creativity, enrichment, and inspiration.

• Write a program to screen scrape the election return data from Dave Leip’s Atlas of U.S. Presidential Elections. Pay careful attention to name clashes between Dave Leip’s site and the U.S. Census (e.g., LaSalle vs. La Salle, Kings County vs. Brooklyn).
• Modify your program to include Hawaii and Alaska.
• Use a map projection (such as Mercator, azimuthal, Albers, or Gall-Peters) to transform longitude and latitude coordinates into points in the plane.
• Explore a different color palette (with 5-7 color categories) for coloring the subregions. Here is one example.
• Write the state name in the appropriate place. For large states, draw it in the centroid of the polygon describing the state.
• Create an interactive GUI which displays the election returns for a county as the user hovers over it. You will probably need to add a method to your polygon data type to determine whether a point is inside the polygon.
• Visualize the gradient or change in votes from one election to the next.
• Visualize a different data set by county, e.g., poverty rate, access to Internet, and average price of health care. Or collect data for elections in another country.

This assignment was created by Kevin Wayne, with inspiration from Bob Vanderbei.