How you talk on Twitter can tell the world how much money you make

September 30, 2015, 7:22 PM UTC
(FILES) File photo dated September 11, 2013 shows the logo of the social networking website 'Twitter' displayed on a computer screen in London. The San Francisco company Twitter announced on September 12, 2013, in a tweet, that it has submitted papers for a stock offering, the most hotly anticipated in the tech sector since Facebook's last year. "We've confidentially submitted an S-1 to the SEC for a planned IPO. This Tweet does not constitute an offer of any securities for sale," the company tweeted. Talk of an initial public offering (IPO) has circulated about Twitter for some time, and the Wall Street Journal estimated the company founded in 2006 is worth some $10 billion. Twitter has become one of the fastest-growing and most influential social media services, used widely by celebrities, journalists, politicians and others. AFP PHOTO / LEON NEAL (Photo credit should read LEON NEAL/AFP/Getty Images)
Photograph by Leon Neal — AFP/Getty Images

New research confirms that all those Twitter comments about the global economy and humble-brags about your nonprofit donations can let people know you’re making a lot of money, assuming that wasn’t your intention all along. Want people to think you’re just an everyday Joe instead? Start swearing more and sounding … wait for it … less afraid.

A recent experiment, led by a researcher from the University of Pennsylvania, analyzed more than 10 million tweets from more 5,100 Twitter (TWTR) users in the United Kingdom to reach this conclusion. The team broke users into segments based on the jobs they listed in their bios, and then calculated the average income for those professions. After that, natural-language processing algorithms analyzed the language in those users’ tweets for the types of words and language most distinct to each income level.

The experiment found that certain topics, as well as sentiments, can accurately predict income a Twitter user’s income level. Users with more money tend to tweet more about politics, nonprofits and corporations, but they also tend to sound more fearful. Lower-income users, on the other hand, tend to swear more.


While the results might seem to reinforce classist stereotypes, the researchers suggest the findings are actually a function of how people use Twitter. Whereas higher income people tend to use it to share news and for professional conversations, lower-income users use Twitter more to more personal chatting with friends. Those types of conversations traditionally involve more informal language.

However, there is a bigger picture far beyond the scope of this study: Social media is an increasingly powerful source for collecting relatively unfiltered data about how people feel and what they believe. This is the reason why so many companies now view Twitter as a new type of de facto focus group, and why large enterprise technology vendors like IBM partner with Twitter to make its data available to corporate customers.

This is also why the President Obama is calling for an increased focus on ethics in data analysis. Because while what we post online might lead to better consumer goods or television shows, they also create new ways to discriminate and invade privacy as companies or even cybercriminals learn more about who we really are.

To learn more about the data privacy, watch this Fortune video:

Sign up for Data Sheet, Fortune’s daily newsletter about the business of technology.

Read More

Artificial IntelligenceCryptocurrencyMetaverseCybersecurityTech Forward