Database Reference
In-Depth Information
source, broadly discussed in the topic, is the GPS (Global Positioning System)
data. We have seen that there is a large variety of devices dealing with this kind of
data: mobile GPS navigation systems, GPS loggers, GPS anti-theft systems, GPS
units for photo cameras, and so on. Clearly, the final use may differ, but these
data sets have some features in common: they all take the global coordinates
(latitude, longitude, and time) of the device and store it for a specific purpose.
Most of them (loggers and navigators, for example) take the information at
regular intervals of time spanning from one second to a few hours, depending on
the final application (e.g., turn-by-turn navigation systems continuously collect
and process the GPS signal), and store it for the final purpose. As we have seen
also in Chapters
1
,
2
,and
3
, a single line extracted from these data typically
includes at least the following information:
ID, timestamp, latitude, longitude, quality of signal
where
ID
is the device id,
timestamp
is the current time, usually expressed
in seconds since the 1970 (higher resolutions may be needed depending on the
application),
latitude
and
longitude
are the GPS spatial coordinates, and
quality of signal
may give information on the accuracy of the measure-
ment. Depending on the application, a user may personally produce small to
large amount of data of this kind. Also, different sources of this kind of data are
available, most of which are not public: anti-theft systems, GPS loggers, and
navigators, for example, are meant to be for personal use, and generally these
data are not publicly available unless the owners explcitly share them on some
social media. Another different source of data is the GSM CDR (Call Details
Record): when placing mobile phone calls, the users generate a large amount of
data about their calls, such as the number called, time, and duration. As seen in
Chapter
2
, a single record of these data has usually the following format:
callerID, receiverID, time, antennaID, start, stop, callID
where
antennaID
is the ID of the GSM base station the phone is attached to in
that moment, and
callID
is used to track the call through different antennas in
the case of a user moving in space. As we see, the spatial information of these
data is much less precise: while the GPS can be accurate to the centimeter, in
GSM data we can only use the antenna ID as geographic information, and this
is very rough. In fact, a single antenna can cover a round region of very diverse
radii, depending on the power of the antenna, the placement of it (within the
city or countryside), and other factors. Therefore, the position of the caller is
estimated with a precision that usually is on the order of hundreds of meters.
Moreover, this is clearly privacy-sensitive information, and these kinds of data
are usually not publicly available. Lastly, there are the data coming from the
online geo-social networks and services. These data are of a different form with
respect to the previous two types as, in addition to the potential geographic
