Thanks, Big Data: America's Drinking Habits Predict the Election
Forget about which candidate has the most likes on Facebook or followers on Twitter. Who needs social networks when "big data" can predict the next U.S. president based on America's drinking habits? That's exactly what we did at LatentView.
Instead of calling likely voters to ask which presidential candidate they favour, our analysts looked at the strong correlation between alcohol consumption and voting patterns.
We built a sweet predictive cocktail, using one part drinking patterns from each state, sprinkled with demographic and historical voting patterns from 2000, 2004 and 2008, and a shot of census data. This also included ethnic and economic profiles adding into our logistic regression model. We then assigned states probability scores to see if voters are more likely to vote for the Democrat or Republican presidential candidate.
We found that Democrats prefer pubs, while Republicans tend to be a bit more reserved and prefer drinking in the privacy of their homes; states with more breweries tend to vote Democratic.
We also observed that the higher the consumption of beer in a particular state, the higher the percentage of Republican votes from that state. Montana, North Dakota and South Dakota rounded things out and landed the top spot with the highest per capita beer consumption. The average North Dakotan resident drank 30.6 gallons of beer in 2011! New York and New Jersey came in last as the most sober states.
The average New Yorker drank just a meager 16.3 gallons of beer in 2011.
Even if Republicans and Democrats can’t agree on much, they both like their beer in equal measure. In the states where republicans are winning, 56% of the total beer is consumed by republicans, and similar results goes for the states in which democrats are winning, with 57% of the total beer being consumed by democrats. With wine and spirits, though, we see some differences. Higher wine (and spirits) consumption correlates to higher Democratic voting percentages.
A high percentage of 58% is claimed by the total wine and spirits consumed by democrats, in states where this party is winning. Republicans rate at 55% in the total wine consumption, in states in which they are winning.
Maximum wine was consumed in Washington DC with the average resident sipping 6.8 gallons of wine in 2011. Least wine was consumed in West Virginia and the average West Virginian took down 0.6 gallons of wine in 2011.
Republicans, surprisingly, love beer in cans, while Democrats fancy their alcohol in glass bottles; 52.5% of all alcohol served in the US in 2011 was served in metal cans, 37.3% in glass bottles and 10% from draught dispensers.
Looking at the demographic and macroeconomic factors alone, Obama woes the most multi- racial voters in Hawaii and the highest household income voters in New York.
However, linking the beer correlation with Republicans and adding this to the predictive cocktail Romney holds sway in states, like Iowa where 88.7% of the population are non-hispanic whites, landing him on top, taking the Presidential win home.
So what does this all exactly mean?
What we’re witnessing in 2012 is the convergence of the big data phenomenon and politics, that’s much more of a skill than merely targeting voters by obvious attributes such as age, gender, geography and website preferences.
Elements such as farmer’s market locations and the less obvious drinking habits are now being evaluated to help give the candidates an added advantage.
This level of micro-targeting might just provide voters with information that is more relevant to their own political priorities.
While Big data can’t yet be used to predict the exact outcome of an election, it’s making things a lot more interesting.
The hottest seat in today’s Presidential campaigns belongs to the Data Mining Analyst.
Who would you rather have a drink with?
Other Posts by Melissa Thermidor
The moderated business community for business intelligence, predictive analytics, and data professionals.
|How do you innovate effectively and maintain a competive edge?|
Learn how in our exlcusive ebook, "Bad Data Need Not Apply: Designing the Modern Data Warehouse Environment."