Information Technology Reference
In-Depth Information
Table. 9.1.
Average number of sunny days in July (1981-2010 normal value).
Observation with interactive extractions
We visualized a dataset containing data values from 1 to 31 December,
2009. Fig. 9.6 shows a zoomed-in view of temperature variation over five
days. Fig. 9.6(a) displays the original view of the tagged temperature data
before applying level-of-detail control, and Fig. 9.6(b) shows the
temperature variation after applying level-of-detail control with tag-based
clustering. Fig. 9.6(a) has many overlaps of polylines, while Fig. 9.6(b)
represents the features of the data more clearly. In Fig. 9. 6(b), we find that
an extreme lower temperature can often be observed from “Clear” or
“Sunny” points, and it exhibits typical features of winter temperature due
to radiative cooling. It is difficult to obtain such knowledge from the
visualization results with shape-based clustering without considering tags,
as shown in Fig. 9.6(c). Here, representative polylines are selected based
only on their shape. In Fig. 9.6(c), polylines which have extremely low
temperatures, and missing important weather features, are drawn as grey.
The result demonstrates that we can observe overall features without
losing detail, while reducing clutter among polylines, by level-of-detail
control considering tags during the clustering step.
Fig. 9.7 shows the overview display when applying level-of-detail
control with tag-based clustering. In the middle of the month (dotted white
rectangle), temperatures drawn as red or yellow polylines decreased
dramatically; temperatures drawn as blue or cyan polylines varied slightly.
To explore long-term variation in the two patterns shown above, we
clicked polylines. Fig. 9.8 (Upper) shows the result when we clicked low-
range points while selecting the two tags “Clear” and “Sunny.” The
resultant polylines tagged as “Clear” or “Sunny” at the clicked points were
highlighted. This shows the range of temperature between daytime and
night to be relatively large. Also, extremely low temperature can be
observed from “Clear” or “Sunny” points. Fig. 9.8 (Lower) shows the
result when we clicked high-range points while selecting two tags “Rainy”
and “Snowy.” Subsequent polylines tagged as “Rainy” or “Snowy” at the
clicked points were highlighted, and indicate that the range of temperature
between daytime and night is relatively small, especially in regions tagged
as “Rainy” or “Snowy,” for long periods in the middle of the month. The
Search WWH ::
Custom Search