Quantified Self
Data Analysis, Design, Protovis
During my graduate program at The School of Information at UM, I took a fantastic course on Information Visualization. For this project I examined my online searching behavior for the previous year, and represented the data in a graphical display. This infovis was built with Protovis, a visualization toolkit for composing custom views of data with simple graphical marks (code.google.com).
February 2010
Search Behavior
March 2009 - February 2010
Roll over for top site per month
Data by Google Web History
Original Data
The following data was taken from my Google Web History. In its raw form, it's almost unreadable and certainly makes it difficult to quickly draw trends and highlights from the data. Visualizing this data in a graphical way, with Protovis, allows for faster data comprehension. And it's just not so ugly!
{month:"January", searches:193, topsite:"vis.stanford.edu/protovis"},
{month:"February", searches:218, topsite:"wikipedia.org"},
{month:"March", searches:162, topsite:"mint.com"},
{month:"April", searches:122, topsite:"tenthousandvillages.com"},
{month:"May", searches:26, topsite:"w3schools.com"},
{month:"June", searches:47, topsite:"menloinnovations.com"},
{month:"July", searches:89, topsite:"wordpress.org"},
{month:"August", searches:142, topsite:"mail.google.com"},
{month:"September", searches:160, topsite:"maps.google.com"},
{month:"October", searches:221, topsite:"wikipedia.org"},
{month:"November", searches:139, topsite:"java.sun.com"},
{month:"December", searches:162, topsite:"wordpress.org"}
