CHRONOVIEW: VISUALIZATION TECHNIQUE FOR LARGE AMOUNTS OF TEMPORAL DATA - Information Visualisation: Techniques, Usability and Evaluation - page 77

a straight line (see Fig. 4.5). In this view, we can see how occurrence times
for the same events differ across a week.
Detailed view
If we click on an event in ChronoView, detailed information about it is
displayed in the sub-window on the right-hand side of the tool. When our
target is tweet data from Twitter, the detailed information includes handle
names, tweet text, tweet date, and so on. We can analyse this data in detail
by observing details of the tweets in this view.
Operation view of the representation
The sub-window on the left-hand side of the tool provides three radio
buttons for switching between the three view types in ChronoView and
sliders for changing some view parameters. By operating the sliders, users
can specify the maximum and minimum frequency of displayed events,
and the opacity of the circles representing events.
If very many events are placed at a similar position, the view becomes
so complicated that it is difficult to distinguish all events at the same time.
In such cases, events can be filtered by occurrence frequency to retain ease
of observation. If very many events with similar occurrence frequencies
are placed at a similar position, filtering by frequency does not work well.
In such cases, adjusting the opacity of the circles representing the events is
effective. The more circles are crowded, the lower the opacity required.
Case study
We illustrate the use of ChronoView in finding human behavioural
patterns by analysing tweet data from Twitter.
Data
We used tweets in Japanese containing the keyword “now”, such as “café,
now” and “Tokyo, now.” “Now” is an English word, but Japanese tweeters
often use it to express their current situation. Tweets containing “now” are
considered signs of casual, daily behaviour. Therefore, analysis of such
data is expected to be useful in the field of market research.