'

Tableau details its natural language query plans

A couple weeks after acquiring ClearGraph, Tableau maps out what natural language query will look like and how it will fit with its machine learning roadmap.

tableau-logo.png

Conversational interfaces with computers have been the talk of tech since the days of Star Trek. Mostly associated with voice response, frustrating experiences interacting with Siri, chatbots, or the interactive voice response (IVR) systems of call centers reveal what a long slog it's been for getting computers to understand natural language, regardless of whether it's in the form of voice or text.

But it took the Amazon Echo's Alexa, which was designed as a conversational voice to Amazon's retail and entertainment services, to show that natural language interfaces could actually perform useful services. When we saw SAS founder Dr. James Goodnight demonstrate how Alexa could be used to query SAS Visual Analytics, we thought that was pretty cool. But when you look at this video, you'll realize that Alexa has only been taught a few things and has a long way to go before it will replace your keyboard or touchpad.

With its acquisition of ClearGraph, Tableau is tackling the more bounded problem of translating natural language to SQL queries. On an analyst call last week, Tableau provided more details of its plans for ClearGraph. It comes atop enhancements that Tableau is already making embedding machine learning to provide a more guided analytics experience. In the latest release, 10.3, Tableau added recommended tables and joins, so instead of being confronted with an endless list of tables, the recommendation engine filters them to the most relevant. In the next version, 10.4, Tableau will start recommending data sources as well. In the future, recommendations for visualizations are likely to be added.

Tableau is not alone in applying machine learning to provide a more guided experience for users. For instance, SAP BusinessObjects offers a guided experience that picks the right predictive analytics algorithms; IBM Watson Analytics provides a guided experience that helps frame the problem and narrative, while Amazon QuickSight guides the user through ad hoc queries.

But for Tableau, the significance here is how that machine learning will be married to natural language query. The goal is enabling the business user to type a question such as "Give me the latest sales results," and the system would convert that to a SQL query to numbers from the latest reporting period. The business user would not have to specify which database, report, or field to get the answer.

The building blocks for natural language processing include a semantic data model that is generated from a mix of inputs. It starts with the metadata from data model, which provides a DBA's view of the data structure. It couples that with usage patterns, which shows how people are using the data, and a dictionary of synonyms and aliases.

ClearGraph, a five-person start-up, developed search-based query, but with an AI twist. Rather than base search on keywords, they rely on context via a patent-pending natural language query technology that is optimized for introspecting databases. It is backed by a knowledge graph that deciphers the context of a query. That graph is built based on user interactions. So if a user asks for a report on the latest transactions, it would examine that person's history to deduce their intent. In all likelihood, a user from sales and marketing typically looks for sales transactions, whereas someone from accounting is probably asking for payables and receivables.

For Tableau, natural language simplifies writing queries, while its previous work with machine learning helps users select and manipulate data sources. While not on yet on the roadmap, another piece that ClearGraph brings is an engine that chooses the best way to visualize the data. In fact, framing the visualization is already addressed by rival offerings like Watson Analytics and QuickSight, but it borders on an area where Tableau continues to partner. For instance, Tableau plans to continue partnering with Narrative Sciences, which specializes in building stories, such as automatically explaining what Tableau visualizations are actually saying. Nonetheless, as Tableau wades deeper into AI, the boundary separating what it and its partners do will blur.