All politics is local, the saying goes. And now there’s an app for that.
Irvine, Calif.-based Alteryx, Inc. has made available Web-based or smart phone software that can predict current presidential election voting patterns down to the ZIP Code. It’s available at this link.
The technical capability to achieve this kind of live, constantly updated yet highly localized details on a national scale is based on something called Big Data. Big Data is the convergence of the Internet and information technology. Applied to campaigns at all government levels, it could up-end twenty-first century presidential politics to a far great degree than direct mail transformed politicking back in the 1970s.
Alteryx explains in a news release that to build this application, the company took “…USA Today/Gallup Presidential polling data, Experian household data, U.S. Census data and third-party registered voter information.” An Alteryx spokeswoman says that the confidential credit reporting data from Experian is aggregated so no individual identities were revealed.
Next, Alteryx “… uses statistical and predictive analysis to predict local voter preferences. Once the application is run, a visual report highlights presidential preferences mapped to a zip code or hyper-local drive time analysis from a provided street address in percentages, local demographic information including mean age, income and housing valuations, as well as providing a national map with the latest predicted polling data at the county level.”
The Big Data-based approach, Alterx adds, “allows a campaign to micro-target its efforts for activities on the ground (e.g., where to send canvassers, where to locate signage, even the selection of a school or community center for a candidate visit) to find the actionable areas where the race is likely to 48 to 50.”
By pooling responses over the last several days of a tracking poll (such as Gallup’s and Rasmussen Reports’) there would be a large enough, constantly updated sample, to allow for nearly live local level updates.
“Short-term bumps are spread around, and the basic closeness of a race in a small area isn’t likely to change much with a two to four point bounce in the national polls (A 48 to 50 local race is as close as a 50 to 48 race, even though this change represents a 4 point shift).”
Could a campaign use Big Data to uncover as-it-happens on Election Day voting patterns straight from electronic voting machines? That’s not likely, the Alteryx spokeswoman added. Such information “comes from (typically county) government sources, and county registrar of voters will not allow live feeds out of voting machines.”
Obtaining and analyzing exit poll results in real time on Election Day is another matter. “It could be possible that exit poll information may be used to ascertain if certain groups are not turning out at the expected level (based on a predictive model), and then targeting last minute GOTV efforts to areas high in groups that have underperforming turnout levels,” the spokeswoman says.
This highly sophisticated use of Big Data during Election Day isn’t likely in the short term, she emphasizes.
In the corporate world, Big Data is in its infancy, but it is already making a big impact on the way companies interact with their customers and each other. The campaigns that are first to figure out how to tap Big Data to make their ground games more accurate and responsive will have a true edge in their respective races.
(Part 2 looks at Big Data’s impact on business and privacy.)
(Under the byline C.L. Talmadge, Candace writes and publishes fiction that explores what happen when politics, passion, and piety collide. More on her series at http://www.greenstoneofhealing.com. Follow her on Twitter @StoneScribe)