Advancements in NLP systems enable all people to have natural conversations with data.
Natural language processing (NLP) brings together computer science and linguistics to help computers understand meaning behind human language. Today, business intelligence (BI) vendors are offering a natural language interface to visualizations so that users can interact with their data naturally, asking questions as they think of them without deep knowledge of the BI tool.
Within the context of modern BI, natural language is being applied to support the analytical conversation. Analytical conversation is defined as a human having a conversation with the system about their data. The system leverages context within the conversation to understand the user’s intent behind a query and further the dialogue, creating a more natural conversational experience. For example, when a person has a follow-up question of their data, they don’t have to rephrase the question to dig deeper or clarify an ambiguity. You could request for a BI tool to "Find large earthquakes near California" and then ask a follow-up question "How about near Texas?" without mentioning earthquakes for a second time.
Machine learning enables systems to gain deeper domain knowledge over time based on a company’s data and the types of questions their users ask. "One of the key characteristics of analytical conversation is avoiding dead ends—being able to ask a question, get a result, and pivot off that original question," explains Vidya Setlur, Development Manager on the Natural Language team at Tableau.
Natural language will also allow users to ask questions based on a data visualization: "Let’s say I ask a question from my BI tool about disease outbreaks and get a resulting visualization. I could ask ‘What is that orange spike?’" says Ryan Atallah, Software Engineer at Tableau. "It's a follow-up question, but it's not based on my data. It's based on the encodings of the visualizations." And when an existing visualization doesn’t make sense in the context of the next question, it will offer an alternative.
Natural language represents a paradigm shift in how people ask questions of their data. When people can interact with a visualization as they would a person, it opens up areas of the analytics pipeline that were traditionally reserved for data scientists and advanced analysts. Users aren’t limited by their analytical skillset—only by their own breadth of questions. It also allows advanced users to answer richer questions in less time and to provide more engaging dashboard capabilities to others. As natural language matures across the BI industry, it will break down barriers to analytics adoption across organizations and further embed data into the core of workplace culture.