Automated Visualization Model Selection for Data Representation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users face difficulties in selecting appropriate visualizations for their data in visualization applications, leading to a lengthy process of trial-and-error and potentially ineffective visualizations due to the lack of guidance in selecting and configuring visualization elements.
Innovation Solution
The system automatically generates suggested visualizations based on user-specific and data-specific attributes by loading a trained visualization model, either the user's or a default model, to provide customized visualization options, which can be refined through user interaction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If users manually select and configure visualization elements through trial-and-error, then they can achieve customized visualizations, but the process requires excessive time and user effort
Solution Approach 1:
The system performs self-service by automatically generating suggested visualizations based on the selected dataset without requiring manual user configuration. The visualization generator analyzes dataset attributes and autonomously produces appropriate visualization options, eliminating the trial-and-error process while maintaining customization quality.
Solution Approach 2:
The system performs preliminary action by pre-training visualization models on various dataset types and configurations before actual use. When a user selects data, the pre-trained model immediately generates suggestions without requiring real-time manual adjustment, thus saving time while maintaining operation ease.
2Adaptability or versatility
If no guidance is provided for selecting visualization elements, then users have full freedom in customization, but the resulting visualizations may be ineffective
Solution Approach 1:
The system implements feedback by analyzing the selected dataset's attributes and using this information to generate appropriate visualization suggestions. The model continuously learns from user interactions and dataset characteristics, providing feedback-driven recommendations that maintain both customization freedom and visualization effectiveness.
Solution Approach 2:
The system applies parameter changes by adjusting visualization parameters based on dataset characteristics. The trained model modifies visualization parameters automatically according to the data type, size, and structure, ensuring effective visualizations while preserving user ability to further customize if needed.
3Productivity
If a trained visualization model is used to generate suggested visualizations, then user effort and time are reduced, but computing resources are required for model training and loading
Solution Approach 1:
The system performs preliminary action by training visualization models in advance before actual visualization tasks. The pre-trained models are stored and reused, avoiding repeated training computations during user interactions. This approach increases productivity while managing computing resource consumption by performing intensive operations beforehand.
Solution Approach 2:
The system uses copying by creating and storing trained model copies that can be loaded and reused multiple times. Instead of retraining models for each visualization task, the system loads copied model instances, significantly reducing computing resource consumption during actual use while maintaining high productivity.
Data Source
AI summary
A computer-implemented method for automatic generation of data visualizations may include: receiving, from a user, a request to open a document, receiving, from the user, a selection of data for visualization in the new document, determining whether the user has a trained visualization model, upon determining that the user has a trained visualization model, loading the user's trained visualization model, upon determining that the user does not have a trained visualization model, loading a default trained visualization model as the user's trained visualization model, using the user's trained visualization to generate one or more suggested visualizations of the selected data, and displaying the one or more suggested visualizations to the user.