Database Reference
In-Depth Information
When the user hits the space bar for the second time, a drop occurs in the view
under the cursor. If the view under the mouse pointer is empty, the software
creates a new scatterplot with the selected data. If the user presses the space bar
while moving over a view-containing data, FromDaDy adds the selected data
to this scatterplot. Although it resembles a regular drag and drop operation, we
prefer to use the term “pick and drop,” because the data are removed from the
previous view and attached to the cursor even if the space bar is released. The
user can also destroy a view if the brush selects all the trajectories and the user
picks them.
12.4.5 Brush Pick and Drop
The fundamentally new aspect of FromDaDy, compared with existing visualiza-
tion systems, is to enable users to spread data across views. Within FromDaDy,
there is a single line displayed per trajectory: trajectories are not duplicated, but
are spread across views. The advantage of this technique is multifold. It enables
the user to remove data from a view (and drop it on to the destination view). The
fly-over view enables the user to rapidly decide if the revealed data (previously
hidden by the picked data) are interesting. Second, it makes it possible to build a
data subset incrementally. In this case, the user can immediately assess the qual-
ity of the selection by seeing it in the “fly-over” view. Furthermore, by removing
data from the first view, the user makes it less cluttered, and this makes it easier
for him or her to pick and drop more trajectories.
Another advantage of the brush, pick, and drop paradigm is that this inter-
action helps the user to perform complex Boolean operations: “I want the tra-
jectories that go into this area but not the ones that are too high and only those
that are faster than a given minimum speed.” A seminal previous work uses con-
tainers (also called layers) to cluster trajectories and explicitly applies Boolean
operations to combine them. Even with an astute interface, Boolean operations
are cumbersome to produce, because results are difficult to foresee. FromDaDy
overcomes this drawback, since all the operations of the interaction paradigm
(brush, pick, and drop) implicitly performBoolean operations. Removing trajec-
tories corresponds to an XOR operation and dropping trajectories corresponds
to an ADD operation. The following examples illustrate the union (AND), inter-
section (OR), and negation (NOT) Boolean operations. With these three basic
operations the user can perform all kinds of Boolean operations: AND, OR,
NOT, XOR, and so on.
In Figure
12.4
, users want to select trajectories that pass through region A or
through region B. They just have to brush the two desired regions and pick/drop
the selected tracks into a new view. The resulting view contains their query, and
the previous view contains the negation of the query.
