In 2003, Stanford University PhD student Chris Stolte presented a novel approach to data visualisation. Rather than composing graphs after data analysis is complete, Stolte proposed a system where the fundamental principles of data visualisation – such as which types of graph are best suited to portray which datasets – are built into that analysis.
To put the technology in to practice, in 2004 Stolte and his PhD supervisor Pat Hanrahan, (who had previously cofounded digital animation company Pixar) started Tableau Software.
Tableau’s software, the company claims, can automatically identify the best way to visualise a certain dataset, and use the visual interface to slice and dice the data. This allows non-experts to analyse complex datasets, it says.
Eight years later, Tableau is now the fastest growing supplier in the business intelligence market, according to analyst company Gartner. Sales doubled in 2011, and have risen 80% again in the first half of this year.
According to Jock Mackinlay, Tableau’s director of visual analysis, the company’s roots in graphical visualisation set it apart from the conventional BI suppliers. “Tableau’s technological roots lie in the computer graphics world, rather than the one inhabited by traditional BI vendors,” he says.
The company’s rise has coincided with a shift in the strategic objective of BI projects, he adds. Fifteen years ago, he says, businesses tried to build monolithic data warehouses to provide a ‘single version of the truth’. “People now realise that that doesn’t exist,” he says.
By allowing more employees to analyse data, Tableau engenders a more participatory approach to business intelligence, says Mackinlay. “In our view, sometimes the most meaningful conversation you can have with somebody is when you both make interpretations about something that are different.”
For Mackinlay, this is a realisation of a life-long ambition. A veteran of Xerox’s Palo Alto Research Center (PARC), home of the modern day graphical user interface, Mackinlay was motivated to democratise access to information from an early age.
“I’d read 1984, and I was seeing the proliferation of data and worrying about how it would be used by large organisations,” he says. “What I wanted to do was put data in the hands of the people.”
But does this democratistion of BI threaten the IT department’s ability to maintain a consistent view of data across the organisation? Perhaps, says Mackinlay, but this can be a positive.
“Rather than IT being a policeman that somehow decides what the truth is, we believe in an open conversation – and yes, there’s going to be different points of view,” he says.
“But IT loves the idea. Pushing the information out to every single person in the company means they can act more efficiently rather than being this slow, lumbering, hierarchical organisation.”
With ‘bring your own device’ firmly on the agenda, and ‘bring your own applications’ on the horizon, IT departments might do well to consider the consequences of ‘bring your own analysis’, as tools such as Tableau might enable.
