Television cameras and and newspapers have chronicled the so-called Occupy Wall Street protest movement as it has grown into a global phenomenon. But what has it looked like online? Thanks to a start-up called SocialFlow, and tons of Twitter data, we can actually see how the idea propagated through influential people and organizations, and across previously invisible conduits to permeate vast expanses of Twitter’s network.
The first ever use of the #OccupyWallStreet hashtag was in an Adbusters blog post, way back on July 13, according to Gilad Lotan, SocialFlow’s head of research and development. The network of tweeters using the hashtag was small and sparse in the beginning, with no major media entities yet participating in the conversation. Above is network graph showing all tweets containing the #OccupyWallStreet 10 days after the first use of the hashtag. The larger and lighter the node, the more retweets it generated.
On October 13, the day NYPD planned to clean up Zuccotti Park, the original site of the protest, the network of people talking about the Occupy movement, represented in the graphic above, was huge. By then, entities like @HuffingtonPost and individuals like @KeithOlbermann were among the influential participants.
Researchers like Lotan are just beginning to dig deeply into the rich data sets produced every day by the users of online social networks. In the November issue of Technology Review, David Talbot describes how Bluefin Labs mines the public conversations and opinions of Twitter users for insights about entertainment, advertising, and politics.
SocialFlow aims to use data from the Twitter and bit.ly fire hoses to help its clients, which include news organizations like The Economist, and retail brands like Pepsi, better use their online network to capture the attention of their audience—and their potential audience.
To optimize the way that your message spreads, “you really have to understand who is following you, and who tends to give you attention,” said Lotan last week during a panel discussion at MIT. Time of day, the specific terms and language used in the tweet, and the topics raised are important too. Online, as in real life, the bridges between disparate social networks are humans. For a message to spread beyond its author’s network such bridges must first notice it, and then be compelled to re-share it.
For Lotan, single tweets or hashtags that spread virally are welcome deluges of data that can shed light on how and why messages spread beyond their original author’s network. There is no recipe for virality, he says, but there is plenty of room for optimization.