Do tweeters in Texas care about taxes more than tweeters in Tennessee?

Over the last two days, a team of researchers from North­eastern Uni­ver­sity, Har­vard Med­ical School and the Tech­nical Uni­ver­sity of Den­mark tracked the lan­guage of more than 20 mil­lion tweets to cap­ture the polit­ical mood of the Twit­ter­verse before today’s midterm elections.

It’s well known that the state you live in plays a role in deciding what issues you care about,” said Alan Mis­love, an assis­tant pro­fessor of com­puter and infor­ma­tion sci­ence at North­eastern Uni­ver­sity who con­tributed to the study. “Because con­ver­sa­tions on Twitter are public, we can geocode indi­vidual tweets and study where Amer­i­cans are talking about spe­cific issues.”

Ear­lier this year, the team ana­lyzed the lan­guage of 300 mil­lion tweets to mea­sure the col­lec­tive hap­pi­ness of people across the country. They found that tweeters on the West Coast are con­sid­er­ably hap­pier than those on the East Coast.

For this project, researchers built a system to scan tweets in real-​​time for polit­i­cally charged key­words, such as “unem­ploy­ment,” “cli­mate” and “ter­rorism.” They chose words by mining about 2,140 web­sites for all 1,152 can­di­dates for gov­ernor and Congress.

Researchers geo­graph­i­cally rep­re­sented the data using a density-​​equalizing map, in which each region is scaled to rep­re­sent its number of tweets as opposed to its land area.

Twitter Nation appeared most inter­ested in job– and gay rights-​​related issues, and least inter­ested in abor­tion– and veterans-​​related issues, according to a crude visual analysis of the map on Monday afternoon.

But, as Mis­love put it, “We can’t say that we see any par­tic­ular trends just yet.”

In another study, the same group of researchers found a large gap between the most pop­ular polit­ical topics among Demo­c­ratic and Repub­lican can­di­dates, according to a daily word analysis of their websites.

Researchers said Democ­rats expended more text on edu­ca­tion and jobs, while Repub­li­cans devoted more atten­tion to taxes and immi­gra­tion. Nei­ther Demo­c­ratic nor Repub­lican can­di­dates talked about the wars in Afghanistan or Iraq or gay mar­riage, they said.

Researchers mapped their find­ings using a so-​​called “cam­paign word­cloud,” in which the font sizes and colors of key­words such as “house,” “Wash­ington” and “country” reflect the rel­a­tive pop­u­larity of words men­tioned at least once on a website.

Dig­ital traces are a pow­erful, yet simple approach to pro­viding news to a mass audi­ence in the same way that public opinion polls are useful for gauging voters’ interest in a par­tic­ular can­di­date,” said David Lazer, an asso­ciate pro­fessor of polit­ical sci­ence and com­puter sci­ence at North­eastern Uni­ver­sity who con­tributed to the study. “These inter­ac­tive tools are quite useful for con­sumers of news who want to under­stand the dynamics of elec­tions in both tem­poral and geo­graphic ways.”

Other col­lab­o­ra­tors on both projects include Yong-​​Yeol Ahn and Yu-​​Ru Lin, post­doc­toral research asso­ciates at Northeastern’s Center for Com­plex Net­work Research; Jukka-​​Pekka Onnela and J. Niels Rosen­quist, both of Har­vard Med­ical School; and Sune Lehman, an assis­tant pro­fessor of infor­matics and math­e­mat­ical mod­eling, at the Tech­nical Uni­ver­sity of Denmark.

View selected pub­li­ca­tions of David Lazer in IRis, Northeastern’s dig­ital archive.