Richard Florida is a co-founder and editor at large of CityLab and a senior editor at The Atlantic. He is a university professor in the University of Toronto’s School of Cities and Rotman School of Management, and a distinguished fellow at New York University’s Schack Institute of Real Estate.
To identify business connections across cities, researchers gathered tweets and retweets using the hashtags #smallbiz and #entrepreneur.
Entrepreneurship may be key to a city’s economic success, but it’s very hard to get useful data about it. If you look only at rates of business formation, you’ll see lots of small mom-and-pop enterprises. But if you just track startup creation, you only capture businesses in particular sectors of the economy that cluster on the coasts.
That’s why a recent study published in the Annals of the American Association of Geographers uses Twitter data to try to get a better handle on entrepreneurial hot spots across the United States. Researchers Feng Wang and Ross Maciewjewski of Arizona State University and Elizabeth A. Mack of Michigan State University gathered information from the social media site to identify the nation’s leading centers for entrepreneurship and map the entrepreneurial networks that underlie them.
The study collected tweets from people identifying as entrepreneurs and analyzed connections between people who tweeted and retweeted the hashtags #smallbiz and #entrepreneur. With a total of 6,507,506 tweets between January 2013 and June 30, 2014, the researchers pulled together profile information and geocoding for 297,639 unique users. Those users aren’t a perfect accounting of American entrepreneurs: Small-business owners who aren’t on Twitter, don’t tweet in English, or don’t use those hashtags slip through the net. But the study does succeed in revealing how this sampling of self-identified entrepreneurs interacts with each other.
Roughly a quarter of the tweets (1.59 million) contained those hashtags and the study ultimately boiled down their sample of users to 89,914 with traceable county level location information. Using network analytics, community detection, and regression techniques, the study traced the flow links between tweeters and re-tweeters.
The map above charts their Twitter activity about entrepreneurship across United States locations. Note the large clusters of activity across the Boston-New York-Washington corridor, around the Bay Area, Los Angeles, and Seattle on the West Coast, and around Chicago, Dallas and Houston, Atlanta, and Miami, which in the main reflect the nation’s major population centers.
The next chart shows the 10 counties with the most retweets on Twitter related to entrepreneurship, followed by a ranking of the subset of ten counties that most frequently interacted with those main hubs by retweeting.
As the study notes, some of these relationships can be explained by proximity, as New York, San Francisco, and Los Angeles all interact with nearby counties. But other counties, such as the District of Columbia or Suffolk, Massachusetts, have tweets that reach across further distances, particularly to the West Coast.
The study then uses an algorithm to group counties into communities based on the intensity of their Twitter interactions, applying flow maps and filtering to show retweets between counties. It breaks down the data into three broad communities based on their networks of Twitter interactions.
The first Twitter entrepreneurship community has a “distinct southern and eastern orientation.” It features major tech hubs, including Suffolk, Massachusetts (Boston), Fairfax, Virginia (nearby Washington, D.C.); and Cobb, Georgia (Atlanta).
The second community features more transcontinental connections stretched across the East Coast, West Coast, South, and Midwest. The key nodes include Marin, California (north of San Francisco); Snohomish, Washington (north of Seattle); and Norfolk, Massachusetts (south of Boston). Of the key nodes identified by the authors for this community, only Mecklenburg County in North Carolina contains its city, Charlotte, within the top Twitter county.
The third community exhibits three key node counties, according to the authors: Henrico County in Virginia, Williamson County in Tennessee, and San Francisco. But the map also demonstrates the reach of West Coast counties such as Santa Clara, Los Angeles, and the Portland area, which connect with major population hubs across the country in Texas, Illinois, Florida, Virginia, and Pennsylvania.
After outlining those three broader communities, the study winnows down into three city hubs of entrepreneurial Twitter—New York, Seattle, and Philadelphia—demonstrating how entrepreneurial networks can now cast across great distances thanks to the internet. Take a look at those in the GIF below.
The study identifies several key features of these entrepreneurial communities on Twitter. The top-ranked correlating factors to entrepreneurial retweets are airports—a proxy for national and global connectivity (with a correlation of 0.97)—and research universities—a proxy for education and knowledge intensity (0.77). Affluence as measured by per capita income (0.56), and internet connectivity as measured by the presence of broadband providers (0.50) exhibited correlations with entrepreneurial retweets that were statistically significant.
Twitter’s entrepreneurial networks still in many ways still mirror our spatial divides. Despite the internet’s seeming ubiquity, the entrepreneurial social networks follow the trends of agglomeration economies, knowledge hubs, and most prominently, the digital divide. In 2015, the Pew Research Center found that 23 percent of online adults use Twitter. More than 54 percent of Twitter users had some college or higher education and nearly 60 percent (58 percent) reported incomes higher than $50,000. Indeed, the geography of entrepreneurship on Twitter is spatially uneven—just as it is in real life.