Database Reference
In-Depth Information
by the extracted places; and (4) analysis of the aggregated data. Visual analytics
and m-events provide a rich tool box of data transformations and visualizations
which help a human analyst exploring the data.
Third, MOD queries deal efficiently with vary large data sets with theoret-
ically no limitation, and are able to express complex queries (neighborhood,
patterns, aggregations, etc.). Although direct manipulation is easy to use (users
are accustomed to manipulating tangible objects), it does not support automatic
exploration. Furthermore, direct manipulation techniques need to be interactive,
which works against the data size. For instance, FromDaDy can display up to 10
million points with an acceptable frame rate. If more data need to be displayed
or manipulated, new computation techniques need to be developed.
Since our visual analytics process uses m-events (geographic and temporal
events), this tool is not suitable for complex computations such as pattern extrac-
tion. MOD can easily extract patterns, but the user needs to know in advance
what he or she is looking for. MOD systems are not good for data exploration.
As a future work, we plan to break the direct manipulation data set limitation
with new interaction paradigms (more complex Boolean operations). We also
plan to combine MOD, visual analytics, and direct manipulation to explore large
data sets. Visualize a small sample, roughly figure out your query parameters,
issue the query in MOD, validate the results by visual analytics, refine the MOD
query, and so on and so forth.
12.8 Bibliographic Notes
For further reading, we recommend the topic by
Card et al.
(
1999
), which details
the information visualization research area. We also recommend the topic by
Tufte
(
1990
), which contains many remarkable visualization instances. Two
conference proceedings contain many examples of visualizations and interac-
tion techniques. InfoVis: The IEEE Information Visualization Conference (IEEE
Transactions on Visualization and Computer Graphics) contains novel research
ideas and innovative applications in all areas of information visualization. Also,
VAST, the IEEE Conference on Visual Analytics Science and Technology, is the
first international conference dedicated to advances in visual analytics science
and technology. The scope of the conference includes fundamental research
contributions within visual analytics as well as applications of visual analytics,
including applications in science, engineering, medicine, health, media, busi-
ness, social interaction, and security and investigative analysis.
The spatio-temporal pattern predicate was first proposed in
Sakr and Guting
(
2011
). It is demonstrated in
Sakr et al.
(
2011
). We used this demonstration as
the basis of Section
12.6
.
