CloudLines: compact display of event episodes in multiple time-series

Area of Science:

• Data Visualization
• Information Visualization
• Computer Science

Background:

• Modern data visualization faces challenges with large, dynamic event datasets.
• Static visualizations struggle to represent incremental data effectively.
• Need for real-time analysis with historical context in domains like finance and security.

Purpose of the Study:

• To propose an incremental logarithmic time-series technique for efficient data representation.
• To address the challenges of visualizing dynamic, large-scale event data.
• To enable comprehensible, real-time visual analysis with historical context.

Main Methods:

• Developed an incremental logarithmic time-series technique adapting to data arrival rates.
• Implemented a decay function for fading data items based on relevance.
• Introduced a distortion magnifying lens accounting for logarithmic time scale distortions.

Main Results:

• The technique effectively visualizes incremental data streams in limited space.
• Maintained comprehensible representations of changing data content.
• Demonstrated validity on online news stream data across different time frames.

Conclusions:

• The proposed technique offers a robust solution for dynamic event data visualization.
• It balances the need for recent event monitoring with historical context.
• Enhances pattern accessibility and readability at various scales.