One of Tableau’s marketing messages says “answer questions at the speed of thought”. I doubt that very many people would really want to do that, the result being a large number of poorly considered questions. In fact, one of the dangers associated with visual analytics is the ease with which we can jump to conclusions because our visual processing is always looking for patterns and significance, even when there is none. This is a very real problem and is the main work of Daniel Kahneman, the Nobel Prize-winning economist.
So it is interesting to see that Tableau has just acquired Empirical, a company whose technology can automatically find patterns and give estimates of significance for those patterns. So instead of jumping to conclusions because something turns up on a chart, the technology from Empirical will say that a pattern is probably just random noise, or that it has real significance. This is a massive step in the right direction if people and organizations are truly interested in an accurate analysis (and this is not necessarily the case).
Of course, all visual analytics platforms allow us to jump to meaningless conclusions. This is not a problem unique to Tableau. Business managers around the world are currently making bad decisions because they regard the insignificant as significant and the significant as insignificant. Having some inbuilt intelligence to guard against such errors is a necessity and not an option. Human visual processing is just too error prone to be relied upon.
So here is an extract from the press release put out by Tableau (minus as much marketing speak as possible):
Tableau Software announced it has acquired Empirical Systems, a pioneering artificial intelligence startup that originated at the Massachusetts Institute of Technology (MIT) Probabilistic Computing Project. With Empirical Systems’ automated statistical analysis technology integrated into the Tableau platform, Tableau customers will more easily gain insight into their data, without needing to manually build the complex underlying data models that would otherwise be necessary.
“We are thrilled to welcome the Empirical team to Tableau to help us bring AI-powered analysis to the masses and enhance the way people interact with their data,” said Francois Ajenstat, Chief Product Officer at Tableau. “Automatic insight generation will enable people without specialized data science skills to easily spot trends in their data, identify areas for further exploration, test different assumptions, and simulate hypothetical situations. Empirical shares our vision of delivering deeper insights to more people through smart analytics, and we’re eager to bring their tremendously talented team to Tableau.”
Empirical Systems’ Analytics Engine automates robust data modeling that would typically require a trained statistician and allows people to explore, explain, predict and simulate that data in real-time. Rather than having to test all variables manually or limit analysis to predetermined hypotheses, advanced statistical algorithms will enhance the user experience and guide people to relevant insights and trends that could have easily gone unnoticed. This makes it possible for people to solve common data problems – like detecting relationships between variables, uncovering previously unknown factors driving patterns or spikes in the data, and inferring missing values – without requiring them to have data science expertise or do custom statistical modeling.