Visualization Recommendation Models for Dataset-Specific Chart Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional visualization recommendation tools in software are inefficient and labor-intensive, often relying on manually designed heuristics that fail to adapt to diverse datasets and user audiences, leading to suboptimal visualization configurations and user understanding.
Innovation Solution
A visualization recommendation system that uses machine learning techniques to automatically generate recommendation scores for visualization configurations by analyzing meta-features of datasets and configurations, employing wide and deep scoring models to provide multiple scoring options for users, thereby improving the customization and accessibility of visualization tools.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If manually designed heuristics are used for visualization recommendations, then the system can provide some visualization guidance, but it cannot evaluate visualization configurations not described by hand-crafted heuristics and becomes rapidly outdated as data visualization configurations are revised
Solution Approach 1:
The patent replaces manually designed heuristic systems with machine learning models that automatically learn optimal visualization configurations from data. The ML models are trained on datasets containing visualization configurations and their effectiveness, enabling the system to evaluate and recommend visualizations without manual heuristic development. This substitution allows the system to adapt to new visualization configurations automatically as they are developed, eliminating the need for technicians to continuously update hand-crafted heuristics.
2Measurement precision
If users manually analyze features of large datasets to select visualization options, then they can make informed decisions, but it takes a long amount of time and is inefficient
Solution Approach 1:
The patent implements self-service by enabling the system to automatically analyze dataset features and generate visualization recommendations without requiring user intervention. The machine learning models automatically extract relevant features from the input dataset, evaluate multiple visualization configurations, and present recommended visualizations to users. This automation eliminates the time-consuming manual analysis process while maintaining or improving analysis accuracy through the ML models' ability to identify patterns and relationships that may be difficult for humans to detect.
3Ease of operation
If contemporary recommendation tools use hand-crafted heuristics, then they can provide visualization recommendations, but they cannot provide recommendations based on a wide variety of visualization configurations and are insufficient for creating customized graphical content
Solution Approach 1:
The patent achieves universality by designing machine learning models that can evaluate and recommend any visualization configuration within the supported visualization software ecosystem. Unlike hand-crafted heuristics that are limited to specific pre-defined rules, the ML models learn from diverse training data encompassing various visualization types, configurations, and effectiveness metrics. This enables the system to provide recommendations for a wide variety of visualization configurations including those not covered by traditional heuristics, making the tool adaptable to both conventional and innovative visualization approaches.
Data Source
AI summary
A visualization recommendation system generates recommendation scores for multiple visualizations that combine data attributes of a dataset with visualization configurations. The visualization recommendation system maps meta-features of the dataset to a meta-feature space and configuration attributes of the visualization configurations to a configuration space. The visualization recommendation system generates meta-feature vectors that describe the mapped meta-features, and generates configuration attribute sets that describe the attributes of the visualization configurations. The visualization recommendation system applies multiple scoring models to the meta-feature vectors and configuration attribute sets, including a wide scoring model and a deep scoring model. In some cases, the visualization recommendation system trains the multiple scoring models using the meta-feature vectors and configuration attribute sets.