A really interesting academic study, Visualizing a Hit, was published by Shaun Ellis and Tom Engelhardt of Rutgers Graduate School of Communication & Information. It’s findings have already been mined and linked by music and media sites, but they all kind of miss the mark; the process may be more interesting and significant than any of it’s results! Notably, this study was executed in a highly automated way, relying on APIs (Application Programming Interface hooks of various music sites) to apply a wide range of objective measurements to a huge body of music (all hit songs from the past 50 years).

The problems with this novel approach are many and deep – in fact it appears many of the reposters didn’t bother to read the conclusions, which note some obvious problems with the data set, before reposting some observations as fact. Still, the automated collection and analysis of public data is probably the really significant aspect of this study. The visualizations and charts represent trends more effectively than they can reveal absolute truths. This approach will become increasingly common as products shift from atoms to bits, opening ways to look at explore and exploit their potential. We can’t address problems or opportunities we don’t see, while visual representation reveals things we miss. Cool stuff!