Sankey Diagrams

A company named HistoryShots offers some fine diagrams, that show the development of political parties in the U.S. and two historical graphs of the civil war (1861 to 1865) that remind of Sankey diagrams. Prints of these maps can be purchased from their site.

The ‘History of the Confederate Army’ and ‘The History of the Union Army’ how armies were formed, regrouped or significantly lost men in battles. The width of the ‘arrow’ with an implicit direction along the time line (shown at the bottom of the chart) represents the number of men. Additionally, the diagram is split into three main geographical areas (Western Theater, Eastern Theater and Trans Missisipi Theater).

These diagrams are inspired by the famous Minard chart, which shows the loss of men on Napoleon’s march to Moscow and during the retreat.

The other two graphs are in the ‘Political’ section of the HistoryShots webpage, and show the development of political parties in the U.S. from 1730 to 1892 and from 1892 to 2005 respectively. On these graphs the widths of the “bands” don’t represent votes or seats. The popularity of each party is shown as the amplitude on the unlabeled x-axis.

Democrats and Republicans are like sinus and cosinus waves over time, a picture that continues on the 1892 to 2005 map. Or one might also be reminded of the intertwined DNA double helix structure…

Note: I presented a similar diagram for the development of political parties in Slovakia 1990 to 2008 here on the blog before.

Kevin Deegan-Krause, an Associate Professor of Political Science at Wayne State University has contacted me with a fascinating idea and a unique example of a Sankey diagram. He writes: “I accidentally “reinvented” the Sankey wheel quite awhile ago for a rather odd usage: political party ebb and flow in E. Europe.”

This diagram he presented in a recent post on his Pozorblog shows how parliament seats were allocated to the different parties in the Slovakian parliament from 1990 to 2008, and how seats were shuffled during the six elections that took place during that time.

“As with energy, the number of party seats in parliament is a closed system and there are flows from some to others. This is a highly modified usage, of course, as we do not know in a precise way “where” votes go from one election to the next, so we just fudge at the day of election and just start over (or have invisible reallocations).”

Have a look, for example at the brown brown arrow: The HZDS party has clearly been on the loosing track. On the other hand SMER (shown in orange), a party that emerged as a spin-off from SDL (pink) in 2000, could gain a significant number of seats in the 2002 elections, when SDL couldn’t make it into Slovakian parliament any more. Others pretty much remained stable in their number of seats over the whole observation period.

I am not sure if Kevin uses the “official” colors of the parties (wouldn’t mind tossing my vote for the pink party ), but a color clustering would also allow to visualize if left, or right, or liberal, or whatever political direction gets stronger or weaker. It would also be a possibility to see, if there is an overall trend in a certain group of countries, such as Eastern Europe, or the apparent growth of socialism in South America.

I am calling for votes on how to name this special type of Sankey diagram. If the first Sankey diagram was for steam power, why not a power Sankey diagram?