Big data presents fascinating possibilities for technology and business, but according to IBM cloud architect and evangelist Raghavan Srinivas, the traditional approach to big data collection and analysis, hypothesis and testing, is outdated and impractical.
"You really want to analyze data in motion, as it's generated," Srinivas said. "For example, if you're doing traffic, there's no point in taking a historical approach about it."
Why is that important? Because IBM has one of the more advanced approaches to big data, and what the company is interested in isn't so much analyzing patterns and predicting outcomes, he said. It's about creating cognition.
"What you're really aiming to get at is this new sexy thing -- artificial intelligence and cognitive [data]. So you really want data to be cognitive," said Srinivas.
To create advanced artificial intelligence, it's crucial to collect data from everywhere and "knit it all together and make some sense in it," as Srinivas said. That data supplies natural language processing and allows for multiple systems of interaction and interpretation by a variety of audiences that include businesspeople, not just developers and programmers, he said.
Finally, Srinivas said, it's necessary to get all data on the cloud, because "cloud is good for failing fast" and as in any new field, be prepared to fail.