Database Reference
In-Depth Information
of a social network like Foursquare web site.
1
A user checks in at a place
at 2 p.m. The next day, she does the same at 1 p.m., and at 4 p.m. she
checks in at another place. Interpolation between these three spatiotemporal
points will be most likely useless for any application that wants to analyze
the movement of this user. However, an application aimed at analyzing the
presence of people in a given area may find this information useful. Note that
the difference between discrete and continuous trajectories has to do with
the application semantics rather than with the time between two consecutive
trajectory points. For example, if we want to perform a long-term analysis of
the positions of people, then it may be the case that the random check-ins
at Foursquare could be considered a continuous trajectory.
Spatiotemporal databases
or
moving object databases
store and
query the positions of moving objects. For example, a typical query to a
moving object database would be “When will the next train from Rome
arrive?”, which is a query about a moving point. We can also query moving
regions and ask questions such as “At which speed is the Amazon rain forest
shrinking?”. However, these databases do not support analytical queries such
as “Total number of deliveries started in Brussels in the last quarter of
2012” or “Average duration of deliveries by city.” These queries can be
more eciently handled if mobility data are stored in a data warehouse.
Conventional data warehouses can be extended in order to support moving
object data, leading to the concept of
spatiotemporal
or
trajectory data
warehouses
which, in addition to alphanumeric and spatial data, contain
data about the trajectories of moving objects. Trajectories are typically
analyzed in conjunction with other data, for instance, spatial data like
a road network configuration or continuous field data like temperature,
precipitation, or elevation.
Like in Chap.
11
, to support spatiotemporal data we make use of a
collection of data types that capture the evolution over time of base types
and spatial types. We denote these types as temporal, and we study them in
detail in the next section.
12.2 Temporal Types
Temporal types represent values that change in time, for instance, to keep
track of the evolution of the salaries of employees. Conceptually, temporal
types are functions that assign to each instant a value of a particular domain.
They are obtained by applying a constructor
temporal
(
). Hence, a value
of type
temporal(real)
(e.g., representing the evolution of the salary of an
employee) is a continuous function
f
:
instant
·
→
real
.
Search WWH ::
Custom Search