Semantic Lenses for Natural Language Data Visualization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data visualization systems face challenges in supporting natural language interactions for specific business use cases, as different groups within an organization may refer to the same data fields and values differently, and existing systems limit data curation to data source owners who may not understand the specific needs of business groups.
Innovation Solution
The introduction of 'lenses' that allow users to curate metadata, including synonyms and suggested questions, specific to business use cases, which are used to interpret natural language commands and generate tailored data visualizations, thereby improving the effectiveness of natural language interfaces.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If data source owners curate metadata for data sources, then data source management is simplified, but business groups cannot access specialized metadata tailored to their specific needs
Solution Approach 1:
The patent segments metadata curation by introducing lenses that separate business-group-specific metadata from the underlying data source. Each lens contains curated metadata (synonyms, descriptions, suggested questions) tailored to specific business groups, allowing multiple specialized views of the same data source without complicating the core data source management structure.
Solution Approach 2:
The patent introduces lenses as intermediary layers between data sources and business groups. These lenses act as mediators that translate business group-specific terminology and requirements into data source queries, enabling specialized customization without requiring changes to the data source itself or its owner's workload.
2Device complexity
If natural language interfaces use generic data source metadata, then system complexity is reduced, but user understanding and interaction effectiveness deteriorate
Solution Approach 1:
The patent applies local quality by providing customized metadata (synonyms, descriptions, suggested questions) specifically for each business group through lenses, while maintaining the generic data source structure. This allows the natural language interface to use simplified generic metadata at the system level while providing enriched, context-specific metadata locally to each user group, improving interaction effectiveness without proportionally increasing system complexity.
3Adaptability or versatility
If multiple data sources are created for different business groups, then each group gets customized data, but system complexity and maintenance burden increase
Solution Approach 1:
The patent makes lenses multi-functional by allowing a single lens to serve multiple business groups with similar requirements, and allowing lenses to be applied to multiple data sources. This universal approach enables customized data access for different business groups while reusing the same lens metadata across multiple data sources, avoiding the need to create separate data sources for each group and thereby reducing system complexity.
Data Source
AI summary
A computing device receives, from a user of a first user group, input specifying a set of metadata comprising synonyms for data fields and data values of a data source. The computing device determines, in association with a second user group using a natural language interface, permission to access the set of metadata specified by the user of the first user group. The computing device, after determining that the second user group is permitted to access the set of metadata, receives a natural language command directed to the data source. The computing device interprets one or more terms in the command based on the set of metadata, and executes queries to retrieve data from the data source. The computing device generates and displays a data visualization in accordance with the retrieved data.