Database Reference
In-Depth Information
16.3 Trajectory from Geo-Social Web
Users in the social web leave footprints of their movements: they visit real and
virtual places and their movements can be recorded and analyzed. Following the
previous scenarios, we want now answer the following question: What kind of
trajectories can we reconstruct from the geo-social web data?
Data we can commonly retrieve and access from geo-social networks is
punctual and discontinuous. The only exception so far is the GeoLife project, an
experiment carried out by Microsoft research in which 165 users tracked their
GPS trajectories on a social platform. The main reason of discontinuity is not
only related to the localization systems (GPS use low mobile battery duration),
but also to the users' communication behavior on social networks. Generally a
user posts a content when it is important for him or her to share it with others
users or friends. This means that he or she is not interested in communicating in a
continuous way. Moreover, some media are more used in specific circumstances
(i.e., photo sharing/repository during holidays), while others in daily routine
(i.e., check-ins or status updates). Following a single user on a single social
network generates a finite list of spatio-temporal positions, which can be used
to implement a discrete trajectory (see Chapter
1
). This use of discrete position
is in a very early stage and still has several limits. One example is the increasing
popular service calledGoogle Latitude, which allows users to share their location
with friends and add it to their status message in other Google applications. The
history option
(in a beta release at the moment of writing) stores the user's past
locations. The user can access a restricted area where he or she can visualize the
trajectory on Google Maps/Earth and a dashboard showing information such as
trips, frequently visited locations, distance traveled, and time spent in different
places. This application uses raw data to reconstruct the user trajectories and
enrich them with semantic information, as described in Chapter
1
for semantic
trajectories and behaviors. In Figure
16.1
it is possible to see one of the authors'
Google Latitude trajectory from one month of data. As it is possible to see, the
trajectory reconstruction in the social network has some challenging issues such
as, for example, the data acquisition, given that the data can be discontinuous in
time. This is manifest in the figure where long straight lines connect far points
on the map. There is no attempt to connect to road map layer or transportation
means.
Following a single user in his or her daily social networks activity on different
social platforms could help in creating different trajectories or in filling some
gaps with respect to using only one social network source. A very nice example
of a
segmented trajectory
(see Chapter
1
) is shown in Figure
16.2
, extracted
from an advertising of a train WiFi connection. The option to share information
between different social networks, publishing content from one platform to
another platform, is a very recent trend and it has not yet been studied by the
scientific community.
