Second, the regions themselves are of wildly differing sizes, and they differ in a way that is not well-correlated with the magnitudes of the underlying votes. The number of electoral college votes won and the share of votes cast within a state or county are expressed in spatial terms, but ultimately it is the numbers of people within those regions that matter. First, the underlying quantities of interest are only partly spatial. Of electoral college votes it has (which in turn is proportional toįigure 7.1: 2012 US election results maps of different kinds.Įach of these maps shows data for the same event, but the impressions they convey are very different.

Right we see a cartogram, where states are drawn using square tiles,Īnd the number of tiles each state gets is proportional to the number Reflect the population of the county shown. The map in the bottom leftĭistorts the geographical boundaries by squeezing or inflating them to The balance of the vote is close to even. Again, the color scale has no midpoint.įourth is a county-level map with a continuous color gradient fromīlue to red, but that passes through a purple midpoint for areas where Third is aĬounty-level map where the color of red and blue counties is graded by Maps colored red or blue depending on the winner. See, first, a state-level, two-color map where the margin of victoryĬan be high (a darker blue or red) or low (a lighter blue or red). Reading from the top left, From top left we Some other ways of representing data like this.įigure 7.1 shows a series of maps of the 2012 US Geographical Information System (GIS), R can work with geographicalĭata, and ggplot can make choropleth maps. When the spatial units of the map are familiar entities, like theĬountries of the European Union, or states in the US. 8.4 Use theme elements in a substantive wayĬhoropleth maps show geographical regions colored, shaded, or gradedĪccording to some variable.8.3 Change the appearance of plots with themes.

6.1 Show several fits at once, with a legend.

5.6 Understanding scales, guides, and themes.

5.2 Continuous variables by group or category.

4.7 Avoid transformations when necessary.

4.5 Frequency plots the slightly awkward way.

4.2 Grouped data and the “group” aesthetic.

4.1 Colorless green data sleeps furiously.

3.3 Mappings link data to things you see.

2.4 Be patient with R, and with yourself.

2.1 Work in plain text, using RMarkdown.

1.6 Problems of honesty and good judgment.

Lat long visualizer 3d