Tableau to launch visual data-prep software

'Project Maestro' announced at user conference along with 2017 Tableau software roadmap that features natural-language queries, a faster data engine and data-source governance.

Tableau is moving into the data-wrangling business, announcing plans for visual data-preparation software code-named Project Maestro.

The idea is to bring the same sort of "self-service" visualization to the prepping and cleaning of data as they've built for data analysis, Dan Jewett, Tableau vice president of product management, told Tableau's user conference this morning. "Maestro is going to make data preparation a breeze."

The software is expected to be available "later next year."

In a brief demo, Jewett showed visual ways of inspecting, joining and editing data. Results could then be piped into Tableau for analysis.

In the conference's opening session, several company executives laid out a product roadmap for Tableau data analysis through next year, including:

  • Speedier data import and analysis. Tableau's new data engine in the works, based on Hyper technology it acquired earlier this year, should make it significantly faster to import and analyze large data sets with Tableau. A conference demo showed hundreds of thousands of records being imported per second, as well as being visualized in real time as the import process continued. The engine is in part a response to customer feedback that building large data extracts took too long.
  • Natural-language queries. Tableau is aiming for true natural speech, not merely being able to type in questions that require using exact field names and functions. The demo included queries such as "show houses near Ballard," which brought up a map of such homes and their selling prices as well as a slider to alter the radius definition of "near." The system would be able to understand a query such as "Show me the most expensive houses near Ballard last summer" without having to explicitly define "most expensive" as the maximim of price field or "last summer" as a particular date range. Next, a user would be able to simply type "show me the cheapest houses" and the system would understand this was similar to the prior query, without needing to repeat "near Ballard last summer."
  • Data governance that allows organizations to certify trusted content and calculations and share useful calculations created in one workbook with all users on the company's Tableau Server.
  • Recommended analyses so Tableau would suggest visualizations and other analysis based on your data set.
  • Pre-built recommended dashboards from cloud services, including combining data from multiple services. This will start next year with four services: Salesforce, Marketo, Eloqua and Quickbooks.

Tableau's conference continues tomorrow and is available via live stream from Austin at

maestro Tableau

Project Maestro screen shot.