Information Technology Reference
In-Depth Information
information from various datasets including: computer system logs, sensor
measurements, and transactions. These datasets easily grow very large
since they include temporal dimensions. Information visualization is a
useful and effective approach for understanding, analysing, and monitoring
such time-varying datasets.
A polyline chart is one of the most popular representations for time-
varying datasets. Newspapers and Web sites show polyline charts for time-
varying datasets including: stock prices, currencies, sports performance,
and weather measurements such as temperatures. Also, we commonly
draw multiple time-varying values in a single polyline chart so that users
can compare time-varying values. For example, when we want to visually
compare the temperatures at multiple locations, we would draw all of them
in a single polyline chart space. On the other hand, we often deal with
hundreds or thousands of time-varying values in the above-mentioned
fields. We may want to visually compare the time variation of stock prices
of many companies, or temperatures of several locations. It is difficult to
read hundreds or thousands of polylines in a single space.
Recent work has addressed the visualization of such large-scale time-
varying datasets. Wattenberg et al. presented a sketch-based query
interface to search for specific shapes of polylines [1]. Hochheiser et al.
presented Timeboxes and TimeSearcher [2], a gradient- and range-based
query interface for polyline-based time-varying data visualization. Some
works have focused on similarity-based pattern and outlier discovery.
Buono et al. presented a technique to search interactively for similar
patterns [3] as an extension of TimeSearcher, and a similarity-based
forecasting technique [4] to forecast future patterns. Lin et al. presented a
technique to discover non-trivial patterns [5], by clustering a set of time-
varying values and searching for outliers. Wang et al. presented a
technique for important polyline selection [6]. Recently, we presented two
time-varying data visualization techniques, featuring sketch query on a
clustered view [7], and pattern display on a heatmap [8].
Meanwhile, such time-varying data is often associated with other
information: for example, temperatures can be associated with weather,
and stock prices associated with social or economic incidents. Such
information may be tightly correlated with time-varying values. For
example, minimum temperature will be much lower in the morning of
sunny days during the winter due to radiative cooling. On the other hand,
difference between minimum and maximum temperature will be smaller
during cloudy or rainy days. This kind of correlation between time-varying
values and associated information can be useful for purposes such as
retrieval of frequent and outlier patterns. However, it is not always easy to
Search WWH ::
Custom Search