Sunday, November 8, 2009

NEST CHARTING IN ANALYSIS


From records of the Framingham Heart Study, conducted from 1948 to the present in one New England city, Nicholas A. Christakis and James H. Fowler mapped out the connections between 12,000 friends and relatives. Fowler and Christakis used the contact information required to track participants for followup in the heart disease study. The authors found that friends had similar levels of obesity. The explanation could be that we have a tendency to choose friends who are similar to ourselves; or that friends are similarly affected by a shared environment. Another explanation is that what is considered an acceptable weight or portion size might be influenced by friends. This idea, of a social behavior being contagious, is not shared by all researchers.
From a book review at nytimes.com/2009/10/04

Examples of these network maps can be seen on YouTube: Prefuse: A Toolkit for Interactive Information Visualization
Jeffrey Heer, Stuart K. Card, James A. Landay
ACM Human Factors in Computing Systems (CHI), 421-430, 2005
PDF (1.2M) | WMV (32.8M) | YouTube

WE ARE ALL DESCENDANTS OF CHARLEMAGNE


My father always proudly believed he was a descendant of William the Conqueror. This simple diagram below shows that anybody would have to be. This was explained to me by Dr. Joseph Chang at Yale. The illustration shows ten generations divided into two groups. By the 8th generation anyone is either without descendants at all or is an ancestor shared by a current individual.

SOCIAL NETWORK ANALYSIS



On Nov. 20 1983 Scott Boorman, a Yale sociologist, wrote a NYT op ed to show that considerable personal information can be gleaned from just knowing who calls whom. Prof. David Krackhardt at Carnegie Mellon Uni. says that article was ridiculed then. Thirty years later the National Security Agency uses just such a database. Valdis Krebs illustrates the kind of insight that can be gleaned in a very simple example shown here.