Representing Unevenly-Spaced Time Series Data for Visualization and Interactive Exploration (2005)

Loading...
Thumbnail Image

Files

TR_2005-73.pdf (133.34 KB)
No. of downloads: 697

Publication or External Link

Date

2005

Advisor

Citation

DRUM DOI

Abstract

Visualizing time series data is useful to support discovery of relations and patterns in financial, genomic, medical and other applications. In most time series, measurements are equally spaced over time. This paper discusses the challenges for unevenly-spaced time series data and presents four methods to represent them: sampled events, aggregated sampled events, event index and interleaved event index. We developed these methods while studying eBay auction data with TimeSearcher. We describe the advantages, disadvantages, choices for algorithms and parameters, and compare the different methods. Since each method has its advantages, this paper provides guidance for choosing the right combination of methods, algorithms, and parameters to solve a given problem for unevenly-spaced time series. Interaction issues such as screen resolution, response time for dynamic queries, and meaning of the visual display are governed by these decisions.

Notes

Rights