Wednesday, December 2, 2015

How a GOP Candidate is Using Big Data


Since mid-November, New Day for America, the superPAC supporting Governor John Kasich, has volunteers gathering in, New Hampshire and Columbus, Ohio, to call potential voters. Their goal is to reach Carson supporters, whose profiles suggest they might defect to Kasich. Within that group, the volunteers prioritize individuals identified as social anchors-people likely to be influential with other potential Kasich supporters.

That information comes from Applecart, a New York based data company. Rather than track connections on Facebook and other social networks sites, Applecart uses analog sources, yearbooks, church lists, amateur sports team rosters, to map relationships among potential voters across the country. The company's executives say the idea is to help a campaign turn out voters through peer pressure. "We are trying to find the one person who will listen to because they went to high school together", said Sacha Samotin, a 24 year old and one of Applecart's founders.

Samotin and two classmates started the company as undergraduates at the University of Pennsylvania, where they worked as research assistants for John DiIulio, a Professor of Political Science who oversaw the establishment of the White House Office of Faith-Based and Community Initiatives under President George W. Bush. In 2012, President Obama's reelection campaign asked volunteers to share their Facebook friend lists, theorizing that potential voters would be more receptive to pitches from people they knew. Samotin and his friends, Matt Kalmans and Anthony Liveris, tried to figure out how to apply the same idea without relying on social network sites.

Applecart trying to reach older voters who might not be online and likely to vote Republican, pulled data from yearbooks, local club registers, or newspaper obituaries held at local libraries. The company also scours public websites, including those of law firms, which often post employee directories. Then Applecart builds a social graph, treating each likely voter as a node in a network. Nuclear family, extended family, colleagues, fellow club members, and former classmates get different weights, based partly on geographic proximity.

Applecart's social graph is shaping the way fundraising is done for the super PAC, which brought in more than $11 million in the first half of the year. The company's analysis took a list of 130,000 donors who made at least three contributions of $1,000 each to Republican candidates in 2006. They winnowed that using several factors, including whether people had given to centrist Republicans like Kasich. Applecart's program spat out 30,000 donors and then separated them into 5,000 distinct social circles based on the connections found in the data. New Day fundraisers are dialing through these lists, hunting for donors who've given generously to campaigns in the past and are connected to other givers, but who never been formally enlisted to fund-raise for a campaign, called latent bundlers.

As more calls to donors, volunteers, and voter are made, Applecart's statistical models will be refined to reflect which connections are most effective at shaping specific types of political behavior. People might be willing to bring their checkbook to a friend's fundraiser, but it's still an open question whether that same person can persuade friends to go to the polls on Election Day.











NYC Wins When Everyone Can Vote! Michael H. Drucker
Digg! StumbleUpon

No comments: