Jonas told his audience at GigaOM’s Structure:Data conference that he’s been recruiting family members for a number of puzzle experiments over the last few months. “My girlfriend sees her son and his cousins — I see three parallel processing pipelines,” he joked.
Jonas’ experiments included omitting 50 percent of the pieces of a puzzle, introducing pieces from other puzzles to add false positives and throwing duplicates from a second set of the same puzzle into the mix. Turns out, nothing can fool a few dedicated teenagers set on solving a puzzle… not even a slightly evil data analytics engineer.
So what did Jonas and his test subjects learn from this? Context matters, as it gives you better predictions, regardless of whether you deal with pieces of a puzzle or vast amounts of data. Also, moments of deep reflection can be really valuable, and offer new insights on existing data. And finally: Real-time analysis of data offers competitive advantages because participants can respond when it matters most – and that’s valuable to both teenage puzzle players and companies using big data.
Watch the livestream of Structure:Data here.