Stack Graph Visualization for Event Pattern Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
It is challenging to detect patterns in large data sets representing events over time, particularly identifying abnormal patterns from a vast dataset, which is akin to finding a 'needle in a haystack,' and existing techniques do not effectively facilitate visual organization to make these patterns readily identifiable.
Innovation Solution
The implementation of a stack graph system that organizes event data into stack lines, allowing for the visualization of event data over time, with features like stack line clustering, secondary graphs, and filtering options to enhance pattern identification, enabling users to easily cross-correlate event data with secondary data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Difficulty of detecting and measuring
If event data is organized into traditional tabular formats, then data completeness is maintained, but pattern detection becomes difficult
Solution Approach 1:
The patent transforms event data from traditional tabular formats into a graphical dimension with stack lines plotted against time. Each stack line represents a target entity, and events are visualized as markers along these lines. This dimensional transformation enables pattern detection through visual inspection while preserving all event data through interactive exploration capabilities.
Solution Approach 2:
The patent introduces an intermediary graphical representation layer between the raw event data and the user. This stack graph visualization acts as a mediator that preserves complete data information while making patterns detectable through visual clustering, temporal relationships, and interactive filtering mechanisms.
2Loss of information
If all event data is displayed simultaneously, then complete information is available, but identifying abnormal patterns becomes difficult
Solution Approach 1:
The patent segments the complete event data set into individual stack lines, each representing a target entity. This segmentation allows users to view all data while simultaneously focusing on specific entities or time periods. Abnormal patterns become identifiable through visual deviations in specific stack lines against the backdrop of complete data.
Solution Approach 2:
The patent implements dynamic filtering and selection capabilities that allow users to interactively explore the complete data set. Users can filter by time range, entity type, or event characteristics to dynamically focus on subsets of data while maintaining awareness of the complete data context through the visualization framework.
3Difficulty of detecting and measuring
If detailed event data is visualized, then pattern detection capability is enhanced, but system complexity increases
Solution Approach 1:
The patent creates a universal stack graph visualization framework that handles multiple event types, time ranges, and entity categories through a single coherent interface. The system provides multi-functional capabilities including filtering, zooming, entity selection, and pattern highlighting without requiring separate visualization systems for different analysis needs.
Solution Approach 2:
The patent enables dynamic parameter adjustment including time range selection, entity filtering, and event type filtering. These parameter changes allow users to enhance pattern detection capability by focusing on specific aspects of the data while the underlying system maintains complexity management through consistent visualization principles.
4Difficulty of detecting and measuring
If temporal relationships are emphasized in visualization, then pattern identification is improved, but data processing complexity increases
Solution Approach 1:
The patent maps temporal relationships onto a visual dimension where time flows horizontally across the graph. Events are positioned along stack lines according to their temporal occurrence, making temporal patterns immediately visible. This dimensional mapping emphasizes temporal relationships while using efficient algorithms to process and position events based on their time attributes.
Data Source
AI summary
A method and system for drawing a stack graph that includes a timeline and one or more stack lines based on a set of event data. A stack line may be associated with an event target and may include one or more event overlays that represent event objects. In one implementation, event overlays may include a visual characteristic that identifies an event source associated with the event object of the event overlay.