Visualization and exploration of temporal trend relationships in multivariate time-varying data

Area of Science:

• Data Science
• Scientific Visualization
• Time Series Analysis

Background:

• Analyzing large, multivariate time-varying datasets is challenging.
• Understanding temporal trends and their interrelationships is crucial for scientific discovery.

Purpose of the Study:

• To develop a novel algorithm for exploring and visualizing multivariate time-varying data.
• To identify and model trend relationships, including correlations and causal effects, within scientific datasets.

Main Methods:

• Developed the SUBDTW algorithm to detect trend appearance and disappearance in time series.
• Modeled temporal trend relationships using a state machine approach.
• Proposed a linear-time complexity algorithm for extracting trend relationships from large datasets.

Main Results:

• Successfully identified important trend relationships by analyzing variable changes over time and space.
• Demonstrated the algorithm's ability to model temporal relationships as state machines.
• Validated the approach on diverse scientific datasets, showcasing its utility.

Conclusions:

• The new algorithm efficiently extracts and visualizes complex trend relationships in scientific data.
• The SUBDTW algorithm and state machine modeling provide robust methods for time series trend analysis.
• Novel user interfaces enhance exploration and understanding of spatio-temporal data patterns.