California enlists data scientists in election monitoring bid

Register now

(Bloomberg) -- Southern California officials are turning to data scientists for help spotting suspicious trends as voters head to the polls in primaries on Tuesday and cyber experts warn that Russia will seek to meddle in this year’s midterm elections.

As part of a pilot project, the Orange County Registrar of Voters is shipping some of its data to researchers at the California Institute of Technology through a secure pipeline. They’re developing analytic tools that could help election officials review voting data for irregularities. California is one of eight states holding primaries Tuesday and Orange County alone has nearly 1.5 million registered voters.

The aim of the California project is to help local officials pinpoint any “anomalies” in their data, providing “metric-based evaluations” on the integrity of elections, according to Michael Alvarez, a Caltech political science professor. The Caltech team will post its data analysis on an online dashboard.

“If there’s any indication that there’s any sort of meddlesome behavior going on,” then “they’ll know about it as quickly as possible and can act,” Alvarez said in an interview.

Forensic Work

Data scientists see Orange County -- traditionally more conservative and Republican than other parts of the state -- as a rich testing ground because it’s large and has closely contested House races. Some of the Orange County districts include seats to replace retiring Republicans Representatives Ed Royce, chairman of the House Foreign Affairs Committee, and Darrell Issa.

Separately, Caltech’s site will run a live feed during Tuesday’s primaries, of election-related keywords, such as “voter fraud,” appearing on Twitter from across the country. Researchers are also tracking similar words specific to tweets that are geo-located to the county.

In addition, the project will deploy election observers to polling places. The county will also provide researchers with precinct data later in the week for them to perform “statistical forensics” on turnout and other trends.

The U.S. Department of Homeland Security has said Russia targeted 21 states’ voter registration systems in 2016. Alvarez’s team is developing a “voter registration auditing tool,” which will allow election officials to monitor changes in their data.

Sowing Doubt

Neal Kelley, Orange County’s registrar of voters, said in a statement that he’s been searching for a way to check the “health” of the county’s election system. “There is no finish line to this process and Caltech’s expertise -- and third-party review --aids us in this endeavor,” Neal said.

Former Director of National Intelligence James Clapper said President Vladimir Putin’s objective during the 2016 election was to sow doubt in the U.S. and that Moscow’s efforts were significantly ramped up from previous elections.

“I’m not sure he cares too much who wins the midterms,” Clapper said in an interview in Washington. But “I think they’ll continue to exploit the schisms and the polarization in this country.”

Alvarez has been involved in election research since the the 2000 presidential race as part of the Caltech/MIT Voting Technology Project. Funding for the current Caltech project came from the John Randolph Haynes and Dora Haynes Foundation.

Researchers are now honing their tools for the November midterms and will continue to refine them heading into the 2020 presidential elections, which Alvarez expects to be “very contested and to be potentially very divisive.”

“If we can do it in the O.C., we can scale it up and try to do it in other locations,” he said.

Bloomberg News