Trajectories and Their Representations - Mobility Data - page 6

during the whole lifespan of the object. Each record (instant, point, features)
contains the instant of the capture, the 2D or 3D point of the object, and possibly
other features captured by the device (e.g., the instantaneous speed, accelera-
tion, direction, and rotation). There are no two records with the same instant
value.
Before going into a detailed analysis of what trajectories are and how they can
be tailored into useful information for the targeted applications, we informally
sketch an example application scenario that uses trajectories to describe the
movement of tourists visiting Paris.
1.2.1 Tourists Application Scenario
Tourism represents an important source of revenue for many countries, regions,
and cities. Its promotion has become a critical business. The efficiency of promo-
tion activities can be boosted by the acquisition of knowledge about the habits
of tourists, their preferences, and the local features that are likely to attract
them in large numbers. Part of this knowledge can nowadays be extracted from
the analysis of on-site movements of tourists, collected via their smartphones
equipped with GPS and connected to social networks.
From a promoter's viewpoint, a tourist destination is a geographical area that
offers tourists the opportunity of visiting a variety of places (e.g., museums,
parks, monuments, and attractions) while using many services (e.g., restaurants,
accommodations, shops, and travel agencies). All these tourist places and ser-
vices are collectively referred to as points of interest (POIs), chosen from a
tourist perspective. A tourist day consists in moving from one POI to another
one, and so on, while stopping for some time in each one of the visited POIs for
eating, resting, shopping, visiting, sleeping, attending a show, or meeting other
people, as shown in Figure 1.1 .
The oriented line in Figure 1.1 shows the spatial route of the trajectory
made by a tourist during one day while visiting Paris. Very often, applications
use only this spatial representation of movement on a background map. It is
very intuitive, yet it provides very little temporal information. Time is only
implicitly conveyed by the fact that the sequence of points forming the line is a
temporally ordered sequence. In other words, going further down the line (from
its beginning to its end) corresponds to moving later in time. In Figure 1.2 , part
of this trajectory is shown with a volume ( x,y,t ) visualization. The trajectory
is represented by the thick line in the upper part of the figure, and its projection
on the ( x,y ) plane shows its spatial route as a line lying on the map. As time
never stops and always flows on, no two points can have the same time value,
and the thick 3D line always moves further on the time axis. When a moving
object stops, its position in the ( x,y ) plan does not change. In the ( x,y,t )