Mapping (some of) the “Tech for Good” Conversation(s)
Marc A. Smith likes to say that he practices “crowd photography for the cyber-square.” That is, as the chief social scientist at the Social Media Research Foundation, he studies computer-mediated collective interaction. One place where he focuses his attention is on the social relationships that you can discover among Twitter users, using a network mapping tool called NodeXL that works as a free extension on Excel. The social graphs that Smith makes show the connections between people as they congregate around an idea and/or an event. Years ago, he built me some images mapping the conversation around the 2012 Personal Democracy Forum, and also looking at the following around various surrogates for the Obama and Romney presidential campaigns.
The other day, in response to a recent edition of Civicist’s First Post that looked at the relationship between “Civic Tech” and “Public Interest Tech,” Smith sent me three network maps, built by looking at the conversational crowd centered on those two terms and a third one, #ResponsibleTech. (If you aren’t familiar with that phrase, it comes from the UK group DotEveryone, which launched a big push for “Responsible Tech” back in 2018.)
Here’s the #CivicTech map. (A more detailed interactive version with lots of additional data can be found here.)
Here’s the #PublicInterestTech map (more detailed interactive version here).
And here’s the #ResponsibleTech map (more detailed interactive version here).
The graphs for #PublicInterestTech and #ResponsibleTech were built from, respectively, 867 and 989 Twitter users who used the relevant term sometime between January 1 and March 12 of this year. The #CivicTech graph is of the network of 3,068 Twitter users who mentioned that term between February 9 and March 12. Those variations in volume make sense, by the way: Civic tech has been around much longer and thus has a correspondingly larger range of participants.
Smith explains how NodeXL constructs these graphs as follows:
These networks are composed of Twitter user accounts that used a particular hashtag. These users are connected to one another if one user mentions, replies to or retweets another. NodeXL draws a line between people who are connected. The line gets thicker when the connections are more frequent. …Users are clustered together into groups if they have lots of connection to each other, based on the frequency of who replies to whom, and who mentions whom. Each cluster is given its own rectangle which has a size in proportion to its population. After they have been grouped in this way, the text of the tweets from the people in each cluster are then analyzed for frequently used hashtags, words and URLs listed in order of use.
Each of the three maps has a rectangle in the top left corner that shows the largest cluster of Twitter accounts for that hashtag. At the center of the big #CivicTech cluster is @CodeForAmerica; at the center of the #PublicInterestTech cluster is the @FordFoundation; and at the center of the #ResponsibleTech cluster is @DotEveryone.
Smith notes that each of the three networks mapped have a “broadcast” structure, meaning that most participants are grouped around a central hub, or a handful of central influencers (you can find more details by clicking through to each of the detailed maps linked to above and scrolling down). He also offers four main takeaways from these snapshots:
• These three hashtags represent three distinct groups of people.
• These groups of people are talking about and linking to different resources.
• These terms may be related but they indicate the existence of distinct populations.
• Building connections among the top ten people in each graph could have the effect of integrating these distinct sub-groups.
Those observations feel spot on. While all three of these concepts are closely related, the main organizational drivers are focused on different tracks of activity. For the time period in question, the Responsible Tech conversation was focused on tools and licenses for responsible innovation, along with a fair amount of attention to ethics in artificial intelligence. The Public Interest Tech conversation was focused, not surprisingly, on the new Public Interest Technology University Network, along with a lot of discussion of public interest tech at a new day-long track at the annual RSA conference. And the Civic Tech conversation was spread across such things as a GovTech article about new developments in the field, the annual BetaNYC School of Data conference, and a French blog post about startups in the fields of civic tech, mobility, edtech, foodtech and green tech. (Civic Tech is big in France.)
Smith’s network maps have one additional intriguing element. In the bottom right corner, there are individual accounts or small pairings where people are using the primary terms in question, but are not connected up to the hubs of conversation. Smith calls these “isolates,” because they are people who don’t connect to any others in the same network. He says, ” A relatively large number of ‘isolates’ in a social movement topic is a good sign, suggesting that a term reaches beyond its core audience of ‘usual suspects.’ A large fraction of isolated users in a topic network is an indication that the topic is not a completely ‘built-out’ or ‘in-group’ network. In other words, isolates are an indicator that there is room for growth based on the casual interest of peripheral participants. Engaging peripheral users is a way to build community!”
So, two big takeaways for everyone engaged in these communities of interest. First, we should look for more ways to cross-pollinate. And second, we all have room for growth.