Sign In

Communications of the ACM

ACM TechNews

Natural Language Interface for Data Visualization Debuts


View as: Print Mobile App Share: Send by email Share on reddit Share on StumbleUpon Share on Hacker News Share on Tweeter Share on Facebook
A data visualization of Antonio Vivaldi's "The Four Seasons."

Researchers at the New York University (NYU) Tandon School of Engineering's Visualization and Data Analytics (VIDA) lab have developed a framework that allows non-experts in machine learning to create highly flexible data visualizations from almost any data.

Credit: Nicholas Rougeux

Researchers at the New York University (NYU) Tandon School of Engineering's Visualization and Data Analytics (VIDA) lab have developed VisFlow, a framework that those who are not experts in machine learning can use to create highly flexible data visualizations from almost any data.

In addition, the researchers made it easier and more intuitive to edit these models by developing an extension of VisFlow called FlowSense, which allows users to synthesize data exploration pipelines through a natural language interface.

Said NYU researcher Bowen Yu, "We believe that with natural language support, we can mitigate the learning curve for a system like this and make dataflow more accessible."

From New York University Tandon School of Engineering
View Full Article

 

Abstracts Copyright © 2019 SmithBucklin, Washington, DC, USA


 

No entries found