Penn map project examines well-being through 37 billion tweets

It’s no secret that communities across the United States differ greatly. Researchers from Penn’s World Well-Being Project sought a simple way to capture, explore, and share such differences on a large scale. Their end goal: to provide individuals with valuable insights about where they live and offer comparisons to other communities.

The result is the Well-Being Map, an interactive, freely available tool based on the statistical language analysis of more than 37 billion publicly shared, geo-tagged tweets and on regional demographic data. Funding for the map was provided by a grant from the Templeton Religion Trust.

For every United States county, the map displays scores for a range of characteristics: well-being and personality traits like openness and extraversion, for example, and also government-reported health and socioeconomic factors such as heart-disease mortality and unemployment. Users can select a single trait to explore each county’s measurements for that trait, or select two traits at once to compare how the traits differ across counties. Top 10 lists show the highest- and lowest-scoring counties in the nation, and within every state for each characteristic. A County Profiles page allows users to enter a county name and see how it ranks in the nation for all reported characteristics.

If the project’s scope sounds massive, that’s because it was more than five years in the making and the long-term goal of the World Well-Being Project, or WWBP itself, says Johannes Eichstaedt, a Penn postdoctoral fellow who co-founded the WWBP, which is part of the University’s Positive Psychology Center, and led the map’s creation.

“The World Well-Being Project was established on the idea that we can measure psychological states of large populations in real time by analyzing their social media content,” he says. “To get here, we collected huge data sets, built language-based prediction models, ran those models over tweets and demographic information, and extracted language patterns associated with specific traits.”

In the past half-decade, the 20 or so Penn scientists involved with the WWBP have continued to promote significant research findings that come out of evaluating social media posts. But founding project member Lyle Ungar, a Penn professor with appointments in the Perelman School of Medicine, the School of Arts & Sciences, and the Wharton School, says he wants the map to reach an even broader audience than the academic work alone.

“I hope people will just want to play around with it. If I were considering moving somewhere in the U.S., I would want to know how happy people are there,” Ungar says. “County-level psychological profiles are really, really important. What people are like in different areas affects economic and health outcomes as well as happiness.”

Christie Versagli, WWBP’s web development director who designed the map’s web interface, says, ideally, the tool would influence policymaking, too.

Digging deeper within the map’s data could also uncover “personality profiles” of communities, identifying places that need additional support, says Anneke Buffone, the WWBP’s lead research scientist.

“We’re looking for areas where we, as a country, can do better,” she says. “By analyzing social media, we can get unexpected, in-depth insights into the processes that drive happiness and illness.”

Though the map is now up and running, the WWBP team plans to keep adding enhancements. The ultimate goal is a live version showing real-time assessments of Tweets as they occur, especially during major events like a presidential election, the Super Bowl, or a natural disaster.

“This event is happening right now. Everyone’s tweeting about it. How are they feeling about it? Is one region responding differently than another? The map could reveal the immediate impact of an event in several different ways,” Versagli says. “This is the type of modern measurement we’re moving toward.”

Staying true to its name, the World Well-Being Project plans to expand the map beyond the United States, partnering with groups in Spain, Mexico, the United Kingdom, and China to uncover how social media can reveal well-being findings across the globe. Eichstaedt says he is also using the map’s information to delve further into his own research.

“We are looking around in the happy U.S. counties and comparing them to happy areas in other parts of the world, to understand how happiness unfolds across different cultures,” he says.

World Well-Being Map