In the 2012 U.S. presidential election, data scientist Nate Silver -- founder of the Fivethirtyeight blog, then associated with the New York Times, predicted the results all 50 states correctly. This brought him national prominence, his achievement hailed as a “triumph of the nerds,” convincing proof of how a skilled analyst can coldly run the numbers and arrive at truth.
Curiously, though, one of the few certainties of the Republican primary in 2016 was that Nate Silver and his FiveThirtyEight colleagues were consistently wrong about damn near everything.
How can Big Data be so reliable in 2012 and such garbage in 2016? The problem isn’t the data, but rather how the FiveThirtyEight team used and interpreted that data. Their errors by no means nullify the benefits of data analytics, but they do point to the importance of understanding people as real flesh-and-blood people, not just as data points on a spreadsheet.