Google Car Shows How Overhyped Big Data Works in Real World
NEWS ANALYSIS: The overused term "big data" may fall out of favor even as the Google Car prototype shows how masses of data are constantly analyzed to keep driverless autos from crashing.
Big data, the overhyped term used to described large scale data analysis, is headed for the big sleep. At least that was the impression drawn from the big data panel that convened May 21 at the MIT Sloan CIO Symposium in Cambridge, Mass.
No less authority than Thomas Davenport, the author of Big Data at Work, said the term big data is “not very meaningful” as he led a panel discussion. In place of big data, terms such as e-knowledge and converged data were offered as alternatives. In the end, it appeared that starting from the business problem you were trying to solve and then deciding on the data required to solve that problem made the most sense.
But just as it seems the big-data buzzword is headed for the cliff, the applications being built from assembling vast data streams are appearing in the technology segment. There was no greater example of digital data—meets sensor data—meets the physical world than Google’s May 27 demonstration of its bug-like driverless car.
Actually the car, which somewhat resembled an early Volkswagen Beetle, took driverless to a new level as it abandoned the old reliables of steering wheel, brake pedal and gas pedal. The car could be called up via a smartphone app and, other than sitting in the car, the passengers would not participate in the navigation.
While it could be years (or maybe never) before the Google-pioneered cars are on the roads, the wisdom behind Google’s efforts to take the driver out of driving makes sense. Google made its first fortune by sending digital robots scurrying around the World Wide Web scoping out the digital content, ranking the value of the content and then making the Web destinations available to anyone with a browser.
While the idea of presenting relevant ads along with those results was not really Google’s idea, the relevant ad model has served the company well. The Google driverless car is a further extension of Web crawling, except in this case the car is crawling the physical world and Google is collecting and analyzing the sensor data that will prevent you (in theory) from crashing in the physical world.
Stripping out the steering wheel, brake pedal and gas pedal makes sense to a techie world where trust in digital technology trumps trust in fallible human judgment. The humans do get a big red panic button to hit if disaster appears imminent. I’m guessing Google will figure out a way to make money from being the king of driverless navigation.
The development of the Google car is a good example of using the right, converged data rather than simply filling up a data lake and then trying to figure out what you are going to do with all that data you have assembled.
In the big data panel at MIT, the panelists included data scientist Puneet Batra and Darrell Fernandes, CIO of strategic investment products and data at Fidelity Investments, both of whom agreed that a convergence is due to sweep through the data collection and analysis vendor community. Those vendors currently offer incompatible point products which lack simple integration tools to multiple data sources. It will be the integration of the varied data streams which yields the biggest data payoff.
A company does not need to be engaged in such far-out projects as driverless cars to rethink their use of data and data analysis. The advertising world is currently in the throes of transition as data targeting for mobile applications represents a big opportunity. But this transition also requires data collection and ad deployment faster and more targeted than that which took place on desktop and laptop browsers.
Big industrial companies such as General Electric are now engaged in building applications to make use of all the sensor data that previously was used only sparingly. It is time for the term "big data" to take a big sleep and start thinking about which types of converged data would be most valuable to your organization.
Eric Lundquist is a technology analyst at Ziff Brothers Investments, a private investment firm. Lundquist, who was editor-in-chief at eWEEK (previously PC WEEK) from 1996-2008 authored this article for eWEEK to share his thoughts on technology, products and services. No investment advice is offered in this article. All duties are disclaimed. Lundquist works separately for a private investment firm which may at any time invest in companies whose products are discussed in this article and no disclosure of securities transactions will be made.