Social Network Analysis, Metadata and Business Intelligence: Paul Revere’s Ride Revisited
Originally published March 18, 2014
“One if by land, two if by sea.” Every child in this country eventually comes across this phrase that shines as brightly in American history as the lantern light from the North Church tower that it refers to. Henry Wadsworth Longfellow immortalized Paul Revere’s Ride in his famous poem, but now Ed Snowden’s leaks about the NSA data collection methods and the debate that it sparked has also brought about an interesting series of attempts to revisit Paul Revere, the man, the ride and the events around it, and use them as a learning scenario for the present.
“I have been asked by my superiors to give a brief demonstration of the surprising effectiveness of even the simplest techniques of the new-fangled Social Networke Analysis in the pursuit of those who would seek to undermine the liberty enjoyed by His Majesty’s subjects.”Using the work of Brandeis historian David Hackett Fischer, who had studied the historical event, as well as Paul Revere himself and his affiliations, in extensive detail, Healy developed some adjacency matrices based on the people Revere knew and the clubs or groups to which he belonged. (In a somewhat cute sideline, the mock report in the blog post credits “information gathered by our field agent, David Hackett Fischer,” whom it calls “…an expert and respected field Agent with a broad and deep knowledge of the colonies.”)
It identifies several key clubs or associations of dubious loyalty to the Crown remarking, tongue-in-cheek, “these organizational names sound rather belligerent.” The groupings are: the St. Andrews Lodge, the Loyal Nine, the North Caucus, the Long Room Club, the Tea Party, the Boston Committee and the London Enemies List. Healy then maps memberships and relationships of individuals and organizations creating both a “Person by Groups” and a “Groups by People” table. Using fairly standard SNA techniques – link analysis and centrality measures – he can rate some of the key members of the network. Guess who appears at the top of the list? None other than Paul Revere. He is at the top or among the top five in “betweenness centrality, eigenvector centrality and Bonacich Power Centrality” scores.
In fact, in an interview (see Paul Revere’s Ride, Booknotes, July 17, 1994), Fischer himself makes the case that the importance of Paul Revere went substantially beyond his ride.
“We found 60 other riders who were out that same night, and it seemed that, far from detracting from Paul Revere, they actually made his role more important in that he was more than just a messenger, he was an organizer. He was a man who would get things done. He was a great joiner. He was an associating man. Everybody seemed to know him.”Of course, this whole exercise was motivated by the Snowden/NSA incident and debate. The many allusions in Healy’s post to the fact that “only metadata” was being used or that the presenter of the report to Parliament worked for the “Royal Security Administration (RSA)” were clearly intended to generate additional debate over that incident. Especially when you move to the next logical step, which Healy does not in his post, but Osterholz did at the ACT-IAC ELC: Now that we know that Paul Revere is a key player in a group of likely “conspirators and traitors,” what do we do about it?
This is the crux, of course, because if you were a revisionist historian at Oxford currently studying that period, it might make interesting speculation to suggest that Paul Revere should have been “neutralized” based on the data. This would sound horrendous to Americans today who still bask in the epic history of the Revolutionary War, where a band of upstart colonists and their ragtag army eventually beat the world’s dominant power…albeit with some help from the French.
At the end of his post, Healy is correct in drawing attention to the potential problem of false positives. “I admit,” he avers, “that, in addition to the possibilities for finding something interesting, there may also be the prospect of discovering suggestive but ultimately incorrect or misleading patterns.” And the consequence of a “neutralization” campaign, where the possibility of a false-positive is not negligible, is a big price to pay to in a society where, theoretically, a person is innocent until found guilty.
In all, this is a rather interesting way of eliciting discussion and transparency on the use of business intelligence techniques on these issues.
Recent articles by Dr. Ramon Barquin
Copyright 2004 — 2019. Powell Media, LLC. All rights reserved.
BeyeNETWORK™ is a trademark of Powell Media, LLC