Database Reference
In-Depth Information
4
Trajectory Data Warehouses
Alejandro A. Vaisman and Esteban Zimanyi
4.1 Introduction
In previous chapters we have seen that the usage of location-aware devices
enables the collection of large volumes of trajectory data. Effective analysis of
such data imposes new challenges for their management, while raising oppor-
tunities for discovering behavioral patterns that can be exploited in applications
such as location-based services or traffic control management.
Data warehouses (DW) and online analytical processing (OLAP) have been
successfully used for transforming detailed data into valuable knowledge for
decision-making purposes. Extending DWs for coping with trajectory data,
leading to trajectory data warehouses (TDW), allows us to extract essential
knowledge from raw or semantic trajectories. For example, a TDW can be used
for analyzing the average speed of cars in different urban areas.
Trajectory data in a warehouse must be typically analyzed in conjunction
with other data, for example, to find out the correlation between the speed
of cars and temperature, precipitation, or elevation. In light of these needs,
in this chapter we provide an overall view that integrates trajectory data in a
more general data warehousing framework, which we call
spatio-temporal data
warehousing
.
We start this chapter by introducing in Section
4.2
the notion of data ware-
housing and describing the main elements in a DW architecture. After giving
in Section
4.3
the running example used throughout this chapter, we address
in Section
4.4
spatio-temporal data warehousing, and show that trajectory data
warehouses can be regarded as a particular case of spatio-temporal DW. We
introduce in Section
4.5
continuous fields and show that they enhance the pos-
sibilities of decision making. In Section
4.6
we discuss a representative TDW,
the one proposed by the GeoPKDD project. We conclude in Section
4.7
.
62
