April 19, 2016

Frank Underwood vs. Donald Trump

Frank Underwood vs. Donald Trump

The data scientists at CITYDATA.ai take a look at the presidential matchup that we all wish could happen.

Excited at the prospect of rambunctious contested conventions? Giddy with anticipation as major political parties splinter and disintegrate? Can't resist binge-watching the House of Cards? Then you must read on, for we bring you the ultimate throwdown.

Ruthless pragmatist versus flamboyant narcissist. Democrat versus Republican. Frank Underwood versus Donald Trump.

It's a dream matchup from a pure entertainment perspective: two egotistical outliers squaring off against each other, each as tough as two dollar steak. The race would be dramatic from start to end. At CITYDATA, we have chosen, however, to examine this race from the data driven, political perspective.

Given the lack of available polling information, we used a demographics-based approach to predict the outcome. With the voting data available from primaries in about 2/3 of the nation's counties, we took a look at the key demographics of support each candidate has received.

The graph below shows 14 demographic indicators. Each dot comes from the population weighted average for the demographic shown. This weighted average is taken for each candidate only within the counties that candidate has won so far. Then this weighted average is compared to the national weighted average and scaled by standard deviations.


The farther left the dot, the less support the candidate has received from the demographic. The farther right the dot, the more support. Donald Trump, for instance, has done great with significantly white counties, while he has struggled with economically strong counties (i.e. High Employment Rate).

Now, using what information was available, both through the show House of Cards and the demographics of the Twitter followers of Frank Underwood, we cobbled together an approximation of his support groups as well.

The next graph compares the support of Frank and Donald. Where Underwood fails with the older population (think America Works eliminating social security benefits), Trump succeeds with those yearning to 'Make America Great' again. Where Trump fails with Hispanics (the wall, racist comments, etc.), Frank succeeds with a surprisingly high number of his Twitter followers coming from Latin America (Argentina & Mexico).


The methodology is mixed and approximate, but it gives a baseline we can use to predict a national election. By combing through all U.S. counties, we can see which ones are more similar to Trump's base and which ones are more similar to Underwood's base. By allocating all counties in this way, we can tally up votes, assign state victors and come up with a final electoral score.

In this alternate reality, the political landscape starkly shifts. Texas runs blue, California turns red, up is down, down is up and it's everybody's un-birthday. We end with a newly elected President Underwood, who grabs 305 electoral votes to thump Trump's 233.


“It only takes ten seconds to crush a man’s ambitions.”
- Frank Underwood