CHRONOVIEW: VISUALIZATION TECHNIQUE FOR LARGE AMOUNTS OF TEMPORAL DATA - Information Visualisation: Techniques, Usability and Evaluation - page 70

Circular and spiral-type visualization
The most popular and effective approach to representing the cyclic aspect
of time is to use either a circular or a spiral time axis. Carlis et al. [15]
developed a spiral visualization approach to represent both the continuity
of time and the periodicity of a week, a month, and a year. The technique
allows us to understand the frequency and periodicity of events. Tominski
et al. [16] subsequently improved the existing spiral representation. They
exploited the two-tone pseudo-colouring method by Saito et al. [17] with a
spiral time axis. With the improved technique, quantitative time-series data
that contains some periodicity can be easily analysed. However, using the
technique to compare many events is difficult when the number of spirals
increases. Ring maps have circular time axes and multiple differently sized
rings for multivariate data [18]. They can be used to observe periodical
characteristics in multivariate data.
Misue [19] proposed an Anchored Map technique for representing
bipartite graphs. On an Anchored Map, nodes in one of two sets are placed
on the circumference and nodes in the other set are placed at suitable
positions to represent their relationships to adjacent nodes. Restricting
nodes in one set to the circumference makes it easier to grasp the
relationships among them. ChronoView can be considered a variation of
the Anchored Map technique. However, whereas Anchored Map is a
technique for graph drawing, ChronoView is a technique for visualizing
data with temporal information.
ChronoView
ChronoView is a technique for visualizing many events with one or more
time-stamps. This technique represents each event as a position on a two-
dimensional plane. Fig. 4.1 shows an overview of ChronoView.
Representation of sets of time-stamps
First, we give a position on the circumference of a circle of radius r to
each time-stamp. We give the 12 o'clock position on an analogue clock to
time-stamp t 0 , and consider the position of events in polar coordinates
based on a circle, as on an analogue clock. Moving clockwise, we place all
time-stamps on the circumference according to time elapsed from t 0 .
Suppose U is the set of all time-stamps, the position of the time-stamps is
expressed by the function f 0 : U ĺ R 2 given by