Scientific Discovery Within Data Streams - Ambient Intelligence for Scientific Discovery - page 71

Fig. 4. Physical Architecture of the Data Stream Processing Engine.
performed by the report bean, using input from the monitor bean (and based
on a method of notification and notification address supplied by the setting
bean) via JavaMail. The system currently supports email and SMS text message
notifications, however, as is true for the entire system, new methods may be
plugged into the architecture.
For the purposes of our overarching concept here, we rely on this architecture
to reduce the initial data-stream to a manageable and interesting set, providing
a dynamic corpus of documents for the visualization component.
3 Visualization Component
One goal in exploratory information visualization is to present information so
that users can easily discern patterns. These may reveal trends, relationships,
anomalies, and structure in the data, and could potentially help confirm knowl-
edge or hypotheses. Patterns may also raise unexpected questions leading users
to new insights. The challenge is to create visualizations that enable users to
find patterns quickly and easily. Researchers at the Pacific Northwest National
Laboratory have developed a suite of visualization tools that help the knowledge
worker in these tasks [3]. Here, we present ThemeRiver TM [4] 1 . We shall discuss
1 ThemeRiver is a registered trademark of the Pacific Northwest National Laboratory.