Database Reference
In-Depth Information
and tasks in analyzing movement data. Furthermore, it defines generic classes
of analytical techniques and links the types of tasks to the classes of techniques
that can support fulfilling them. The techniques include visualizations, data
transformations, and computational analysis methods developed in several areas:
visualization and visual analytics, geographic information science, database
research, and data mining.
Readers interested in visualization of trajectories and techniques for inter-
action with the displays can be referred to the papers by
Kapler and Wright
(
2005
) describing a nice implementation of the space-time cube,
Bouvier and
Oates
(
2008
) suggesting original interaction techniques for marking moving
objects on an animated display and tracing their movements, and
Guo et al.
(
2011
) showing the use of several coordinated displays and interactive query
techniques specifically designed for trajectories, such as sketching for finding
trajectories with particular shapes.
Rinzivillo et al.
(
2008
) talk about visually supported progressive clustering of
trajectories. The paper argues for the use of diverse distance functions addressing
different properties of trajectories, describes several distance functions, and
demonstrates the use of progressive clustering by example.
Andrienko et al.
(
2011b
,
c
) refer to “looking inside trajectories” (Section
8.3
).
The first paper describes visual displays that show temporal variation of dynamic
attributes associated with trajectory positions. The second paper gives a struc-
tured list of position-related attributes that can be computationally derived from
movement data alone and from a combination of movement data and contextual
data. These attributes characterize either themovement itself or possible relation-
ships between the moving objects and the movement context. Both papers deal
with extraction of spatial events from movement data. The first paper introduces
a conceptual model where movement is considered as a composition of spatial
events of diverse types and extents in space and time. Spatial and temporal rela-
tions occur between movement events and elements of the spatial and temporal
contexts. The model gives a ground to a generic approach based on extraction
of interesting events from trajectories and treating the events as independent
objects. The paper also describes interactive techniques for extracting events
from trajectories. The second paper focuses more on the use of extracted events
in further analysis. Thus, it considers density-based clustering of movement-
related events, which accounts for their positions in space and time, movement
directions, and, possibly, other attributes. The clustering allows extraction of
meaningful places. The further analysis involves spatio-temporal aggregation of
events or trajectories using the extracted places.
Andrienko and Andrienko
(
2010
) give an illustrated survey of the aggrega-
tionmethods used for movement data and the visualization techniques applicable
to the results of the aggregation. These methods and techniques are also pre-
sented in a more formal way by
Andrienko et al.
(
2011a
).
Willems et al.
(
2009
)
