FPCS: Feature Preserving Compensated Sampling of Streaming Time Series Data.

Area of Science:

• Computer Science
• Data Science
• Information Visualization

Background:

• Data visualization is crucial for intuitive data analysis across various scientific and financial domains.
• Visualizing massive, continuous streaming time series data often faces network pressure, causing lag or rendering failures.
• Existing methods struggle with efficient sampling for dynamic, high-volume data streams.

Purpose of the Study:

• To propose a universal sampling algorithm, Feature Point Compensation Sampling (FPCS), for efficient visualization of streaming time series data.
• To address the challenges of data transmission and network load in real-time data visualization.
• To develop a method that retains essential data features for high-quality visualization.

Main Methods:

• Developed the Feature Point Compensation Sampling (FPCS) algorithm.
• FPCS retains feature points from continuously received streaming time series data.
• The algorithm compensates for fluctuating feature points to ensure accurate representation.

Main Results:

• FPCS optimizes sampling by compensating for feature points, preserving original data visualization characteristics.
• Achieved the shortest execution time compared to existing sampling algorithms.
• Demonstrated negligible space overhead and independence from overall data size.
• Successfully applied to both infinite streaming and finite static data.

Conclusions:

• FPCS offers an effective solution for visualizing large-scale streaming time series data.
• The algorithm provides high-quality sampled data with superior efficiency and minimal resource usage.
• FPCS is a versatile tool applicable to diverse data scenarios, bridging a critical gap in data sampling techniques.