169  |  Data Conversations with Vidya Setlur

2022-12-12 · Source: Data Stories · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics, Software Development & Engineering · Depth: Advanced, extended

Summary

Vidya Setlur, Director of Research at Tableau, discusses the critical role of natural language processing (NLP) in data visualization and analytics. Setlur, co-author of "Functional Aesthetics for Data Visualization," highlights her work on Eviza, a research prototype, and its product counterpart, Ask Data, which enable natural language interaction with data. The discussion covers the challenges of transitioning research prototypes to products, the interplay between research and engineering, and the multifaceted ways language influences visualization, including semantic icon encodings and categorical color assignments. Setlur also touches on the impact of large language models like GPT-3 on data analysis, emphasizing their current limitations in numerical understanding despite advancements in general language comprehension. The conversation underscores the importance of mixed-initiative systems and the nuances of human-computer interaction in data exploration.

Key takeaway

For AI Scientists and Research Scientists developing data visualization tools, you should prioritize integrating natural language processing to create more intuitive and context-aware user experiences. Focus on building mixed-initiative systems that combine the strengths of direct manipulation with the flexibility of natural language, recognizing that users adapt quickly to system capabilities. Consider how language can provide scaffolding and context, especially for complex analytical intents and "why" questions, to enhance data accessibility and interpretation beyond purely visual encodings.

Key insights

Natural language processing significantly enhances data visualization by enabling intuitive interaction and deeper semantic understanding.

Principles

• Human-computer dialogue benefits from pragmatic context and adaptive systems.
• Mixed-initiative systems combine direct manipulation with natural language for optimal user experience.
• Text and annotations are crucial for data interpretation and scaffolding user understanding.

Method

Eviza and Ask Data leverage natural language utterances to connect with Tableau's visql command stack, supporting aggregations, groupings, filters, sorting, and refining vague numerical concepts through user interaction.

In practice

• Use natural language interfaces to query data more intuitively.
• Incorporate text annotations to provide context and support data interpretation.
• Design systems that adapt to user input and remember context for improved interaction.

Topics

• Natural Language Processing
• Data Visualization
• Conversational AI
• Augmented Analytics
• Human-Computer Interaction

Best for: AI Scientist, Research Scientist, AI Researcher, NLP Engineer, Data Scientist

Related on AIssential

• How Machines Understand Human Language: From Traditional NLP to Modern AI — Machine language understanding integrates domain, discourse, and world knowledge with intent inference, moving beyond traditional rule-based NLP.
• INFOFISCUS Conversa | Conversational Analytics Platform for Enterprise Data Democratization — Conversational analytics platforms democratize enterprise data access through natural language querying and AI-powered reporting.
• Building IdeaScape: Turning Thoughts Into Interactive AI-Powered Knowledge Graphs — AI can augment human thinking by visualizing existing patterns in personal knowledge.
• Data is hungry for context — Multimodal AI processing enriches understanding by integrating diverse data types like audio, video, and images.
• When his hobbies went on hiatus, this Kaggler made fighting COVID-19 with data his mission | A… — Combining sentence embeddings with a custom BERT QA model and study design classification enhances medical document search and information extraction.
• An observation on the subway that changed how I think about voice AI — Human speech is a primal interface, making voice interaction a natural evolution for human-computer communication.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Data Stories.