Neighborly Data: Dato to Integrate Machine Learning Services with Tableau
Dato, a Seattle machine learning startup, will put its services in front of a large audience of potential customers through an integration with the next product release from its neighbor, Tableau Software.
Tableau (NYSE: DATA), the data visualization and analytics company headquartered in Seattle’s Fremont neighborhood, is currently beta-testing Tableau 10, which is due out later this summer and will include the Dato Predictive Services integration along with a host of new features.
Dato product manager Roman Schindlauer says the integration will allow Tableau users to create predictive datasets within Tableau using the Python programming language, along with its hundreds of machine learning libraries and tools.
That will enable “more complex scenarios” within Tableau—things like sentiment analysis, churn prediction, lead scoring and other predictive analytics that help companies put the reams of data they gather to good use, he says.
“It’s really the ability to make predictions about the potential future behavior of your users as a company,” he says.
Companies that collect data on usage of their products or services, past sales, or recent customer contacts, for example, can build a Tableau workbook on top of that data and run a Dato machine learning model to predict which customers might be about to switch to a competitor.
“We’re making it very smooth to implement that within Tableau,” Schindlauer says.
That sort of predictive analytics work is typically done now by periodically running a machine learning model over a stand-alone database of user interaction data, then exporting the results to software like Tableau for visualization and analysis. (Dato itself is a developer platform and has no comparable visualization tools built in, Schindlauer says.) It’s a slower, less-efficient way of working that requires access to the data, permissions, and multiple teams within an organization, he says.
Having the capability directly within Tableau will allow analysts without deep machine learning expertise to work independently within the Tableau environment they’re used to, he says.
For Dato, the integration opens up an audience of Tableau’s 42,000 customer accounts—“orders of magnitude more users” than Dato has at this point, Schindlauer says. To use Dato Predictive Services from within Tableau, customers must have it set up independently—either as a cloud service or on premises.
This partnership brings together pieces of what Madrona Venture Group managing director Matt McIlwain has called the dataware technology stack—a framework for thinking about the interlocking technologies that businesses are using to derive value from the promise of big data. (Madrona is a Dato investor.)
The fact that it’s between two companies located within walking distance of each other underscores Seattle’s leadership in this broad field. Both Tableau and Dato also have a mission of democratizing access to data analytics and visualization tools, and machine learning capabilities, respectively.
“They’re very close. Just across the street,” Schindlauer says, adding that the proximity has eased work on the technology integration.
He continues: “I think Tableau has always been a company that Dato looks up to as an extremely successful startup, basically creating a whole new field of democratizing data analysis and data visualization. We at Dato have always looked at Tableau as something that we want to be in a few years, in terms of success and vision.”