Friday, August 31, 2012
Ending August with a final look back at the week in data news.
- Matt Fates of Ascent Venture Partners writing for the Boston Globe talks about how data-driven innovation can boost the economy of Massachusetts, especially given the governor's "Big Data Initiative." Fates explains, though, that extracting the benefits means more than just having "big data"—it's about the innovation and the insight that we derive from it:
In order for magic to happen, Big Data firms need to adopt a different mindset and step into the shoes of the organizations they sell to. They need to not only be able to provide or collect the data, they need to make it relevant to potential enterprise clients and explain why it all matters. They need to help retailers and banks, for example, understand how bits and bytes can make them more responsive to customers or help them adapt to market conditions in real time.
This necessary shift – from being able to develop technology to being able to convince corporations of its value – is essential and especially difficult for entrepreneurs who may have been trained as engineers. The technology is critical, but it must translate into complete solutions to gain adoption.
- Forbes' Dave Einstein wrote about social networks' plans to use content posted by users to predict in real time the outcome of the 2012 US presidential election. But the author cautions that the portion of the electorate who uses social networks is not a representative sample of the electorate at large, likening the split to an example from American Presidential Elections past:
...Social networks tend to skew toward an upscale, younger demographic, and they skirt lower-class neighborhoods like tourists in Las Vegas who never leave the Strip. Put another way, the elderly women in Philadelphia who doesn’t have a photo ID also probably doesn’t tweet much or contribute to the 15 terabytes of new information on Facebook every day.Despite the drawbacks, it's certainly going to provide a really interesting perspective on at least how a subset of the US electorate thinks and feels about the election.
In fact, technology can be a two-edged sword. Remember the 1948 election, when the polls predicted a Dewey landslide over Truman? That election marked the first time pollsters relied on telephone surveys, giving them access to more voters—big data back then. It turned out that a lot of Truman supporters didn’t have phones. A parallel exists today: a growing number of U.S. households rely on cell phones instead of landlines, and pollsters don’t have access to most cell phone numbers.