U.S. Voter Data Gets Wrangled

(Brian A Jackson/Shutterstock)

The unforeseen outcome of the 2016 U.S. presidential election underscored the need for more granular data about the American electorate and its attitudes. Save for a few datasets focusing on national campaign contributions, nearly every pollster missed the populist surge that elected Donald J. Trump.

Hence, data analytics and wrangling tools are emerging to make sense of non-standard U.S. voting records spanning more than 3,000 counties. Among them is a “community organizing” software platform called NationBuilder designed to provide voter registration data to local, state and national political campaigns.

Data wrangling specialist Trifacta said Tuesday (Feb. 28) that NationBuilder has incorporated the tool into its platform to standardize about 145 million voter records from each of the 50 states. The data are then made available free to national and state-level political campaigns.

A data organizing process that used to take two years can be reduced to less than two months, the partners said.

“Finding, requesting and organizing [voter] data is a task campaigns must execute at great cost and effort,” NationBuilder CEO Jim Gilliam noted in a statement announcing the partnership. “The challenge is nearly every region of every state records the data in different ways, making the data messy and difficult to work with.”

The goal is creation of a previously “unheard of” free voter file out of uncategorized voter registration data, added Trifacta CEO Adam Wilson. The tool creates a unified source of voter information that would be available to campaigns that previously had to build the databases from scratch at great expense.

Among the challenges was the requirement for constant updates of large voter datasets. That required NationBuilder to frequently refresh millions of voter records in order to keep up with campaign cycles at the local, state and national levels.

The tool also is touted as allowing campaigns to shift resources from developing expensive voter database to providing non-technical campaign staffers access to tools for wrangling and maintaining voter data from one election to the next.

Political campaigns need to know “who do I need to communicate with? Who needs to vote?” explained Gina Davis, NationBuilder’s vice president of professional services. Hence, the company attempted to create a national voter file that would be “relatively consistent across the country,” Davis added.

“Each state and many hundreds of counties have their own little iterations of the data,” explained Jim Harrold, a NationBuilder data services engineer. “They store it differently,” meaning that compiling a single state file could take weeks.

It took the company two years to complete the first version if its national voter file. When they attempted to refresh the file, “it fell completely over,” Davis said. Hence, the partnership with Trifacta allowed NationBuilders to “translate the logic that we’d already built into a Trifacta Wrangler script that was going to be more useable, more robust.”

San Francisco-based Trifacta began giving away its “Wrangler” tool for Mac and Windows platforms in the fall of 2015. The tool is designed to automate the data-cleansing process in preparation for analysis using visualization tools.

Recent items:

Polling Data Hidden in Plain Sight

Fake News, AI and the Search for the Truth