By Dave Hannon
When you start talking about "data analysis" to the average non-IT person, their eyes glaze over pretty quick. But when you say "mountain lions on treadmills" suddenly you have their attention. Well, you know why that mountain lion is on a treadmill? To collect data.
The real-world day-to-day applications of data analysis are really popping up in new and wildly interesting ways and it's the IT folks -- that's you! -- that make it all possible.
According to this New York Times article, researchers in Colorado are testing data-colleecting collars that can be put on mountain lions to track their movements and biological processes. Interesting you say, but not that practical? Well, like all good data analysis, the researchers coordinating this project hope to use the data to not only track past big cat events, but predict future ones.
According to the article, the data stream from a smart-collared animal would show, for example, when the animal had last eaten, and how likely it was to go beyond its normal range in searching for food.
“We want to get to a stage where we can say, ‘We’ve got a lion that, for whatever reason, is really hungry out there and chances are you should put your dog indoors and shouldn’
t go hiking in this area’ — that there’s a higher likelihood that this animal is going to go after something,” said Terrie Williams, a professor of biology at the University of California, Santa Cruz, and one of three co-investigators on the project.
Cool, you say, but you don't live in mountain lion country. You live in the city and your most pressing problem is getting to work on time and filling up your gas tank, not getting eaten by a massive predatory feline. That's okay. There's an app for that too.
Researchers at MIT and Princeton have developed a smartphone app that uses a network of smartphones mounted on car dashboards to collect information about traffic signals and tell drivers when slowing down could help them avoid waiting at lights. According to this report in MIT News, "By reducing the need to idle and accelerate from a standstill, the system saves gas: In tests conducted in Cambridge, Mass., it helped drivers cut fuel consumption by 20%."
(As an aside, there is probably no better place to test this app than Cambridge. Combine 300-year-old street layouts with thousands of student pedestrians all wearing earbuds, and masses of traffic signals and it takes a half-tank to get from Fresh Pond to the Galleria. But I digress.)
In case you're wondering, the researchers did model a feature which tells you when to speed up to make lights, but decided that was not a safe option to have. (I smell a big market opportunity for a startup with a good legal team!).
These are just a couple real-world examples I came across today that show the migration of data analysis from the business world to Main Street (literally). And these types of applications are going to require lighting fast analysis of unstructured data.
Anyone know a g
ood technology for doing that? Anyone? Bueller? Four letters ... starts with "H"...