Snowy: Recommending Utterances for Conversational Visual Analysis
summary
SNOWY is a prototype that generates and surfaces parameterized natural language utterances for visual analysis by combining data interestingness and language pragmatics. It inspects the current chart, past interactions, and a catalog of possible operations, eliminates inapplicable options, reorders and parameterizes remaining choices, then renders phrased suggestions via grammar rules. A remote preliminary study with 10 participants (50 minute sessions) shows recommendations guide exploratory workflows and teach phrasing even when users question relevance, so SNOWY effectively creates parameterized utterance recommendations that change views and prompt follow‑up analysis.
AI-written summary · about
chapters (4)
related talks
20mEviza: A Natural Language Interface for Visual Analysis
22mDiscovering natural language commands in multimodal interfaces
35mNatural and conversational shopping experiences
24mInferencing underspecified natural language utterances in visual analysis
28mTalking to Objects in Natural Language: Toward Semantic Tools for Exploratory(…)
20mDataTone: Managing Ambiguity in Natural Language Interfaces for Data Visualization
10mFrom Philosophy to Interfaces: an Explanatory Method and a Tool Inspired by Achinstein's Theory ...
16mType-Directed Synthesis of Visualizations from Natural Language Queries