Parallel Coordinate Plot Aggregation for Multivariate Data Visualization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for visually presenting multivariate data using parallel coordinates face issues with visual clutter and overplotting, especially with large data sets, making it difficult to recognize patterns and requiring significant computational resources.
Innovation Solution
The system aggregates individual data points into curves between axes, allowing for the display of large data sets, including hundreds of thousands to billions of points, and enables users to examine relationships between variables by transforming axes, thereby reducing rendering time and improving data visualization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If parallel coordinates plot is used to visualize multivariate data, then relationship between variables can be overviewed, but visual clutter increases for large data sets
Solution Approach 1:
The patent merges multiple individual data point representations into aggregated curve segments. Instead of displaying each data point as a separate line, the system combines collinear segments from multiple data points into single curve segments, reducing visual clutter while preserving the ability to view variable relationships across large data sets
Solution Approach 2:
The system dynamically adjusts the level of aggregation based on the number of data points and available display space. When data sets are large, aggressive aggregation into curves is applied; when data sets are smaller, individual points can be displayed. This dynamic adaptation resolves the contradiction between showing all data points and avoiding visual clutter
2Measurement precision
If individual data points are plotted separately in parallel coordinates, then each data point is visible, but overplotting occurs which hampers pattern recognition
Solution Approach 1:
The patent merges individual data point representations into aggregated curve segments. Collinear segments from multiple data points are combined into single visual elements, eliminating overplotting while maintaining data integrity. This allows patterns to emerge clearly from the aggregated curves rather than being obscured by overlapping individual point lines
Solution Approach 2:
The system transitions from one-dimensional line representations to two-dimensional curve segments with varying thicknesses. The thickness of curve segments encodes additional information about data density, allowing pattern recognition while preserving data point visibility through a dimensional transformation
3Quantity of substance
If large data sets with hundreds of thousands to billions of points are displayed, then comprehensive data analysis is enabled, but rendering time increases significantly
Solution Approach 1:
The patent merges thousands or millions of individual data point renderings into a manageable number of aggregated curve segments. This consolidation dramatically reduces the computational burden of rendering while preserving the ability to analyze comprehensive data sets, as the aggregation process groups collinear segments that would otherwise require separate rendering operations
Solution Approach 2:
The system performs preliminary aggregation of data points into curve segments before the actual rendering process. By pre-processing the data to identify and group collinear segments, the system reduces the complexity of the rendering operation, enabling fast visualization of large data sets without sacrificing comprehensive data analysis capability
Data Source
AI summary
Various systems, methods, and media allow for graphical display of multivariate data in parallel coordinate plots and similar plots for visualizing data for a plurality of variables simultaneously. These systems, methods, and media can aggregate individual data points into curves between axes, significantly improving functioning of computer systems by decreasing the rendering time for such plots. Certain implementations can allow a user to examine the relationship between two or more variables, by displaying the data on non-parallel or other transformed axes.