Database Reference
In-Depth Information
information contained in it, they also typically include the content. We can,
in fact, detect two blocks of information in this type of data: the geo-location
and the media payload, the latter being either a text message, or a picture, a
video, and so on. The content contained in the second block is then said to
be “geo-tagged” according to the first block of information. The geographic
information contained in this kind of data is usually a derivative of the GPS data
coming from mobile devices (e.g., smartphones, PDAs, cameras) from which
the user generated the content. Thus we can still consider this as a GPS data
source, even if, given the very particular features of the application (i.e., no
need to continuously track the user, no need for a specific precision, and so
on), these data are differentiable from the one coming from the GPS navigators
and loggers, in terms of precision, temporal resolution, and final volume of
data collected. Given the large amount of geo-social networks and services
available nowadays, it would be impossible to list all the possible information
available online. We can, however, present here three different types of services,
representative of a large set of available online social networks: Twitter, Flickr,
and Foursquare. Twitter is a social network where users can post short messages
in their timeline (typically publicly available) that will appear automatically
within the timelines of all their followers. A typical message is a text message
no longer than 140 characters, which may contain text and URLs for attached
media such as pictures or videos. The messages can also be geo-tagged if the
user has enabled this feature. A typical piece of data then contains the following
information:
userID, messageID, text, geo-location, timestamp
Flickr is a photo sharing service with a social network layer where users can
post pictures and video in their profile. Tags, comments, geo-location, and EXIF
data (technical data about the picture) are usually associated with the pictures
(or videos). A typical piece of data regarding a picture contains the following:
pictureID, userID, geo-location, timestamp, tags, comments
Foursquare is a location-based social network where users can post their
current location and share this with all their friends. The service includes game
features to incentivize the users to share their location. A typical piece of data
contains the following:
userID, geo-location, locationID, timestamp
Tabl e
16.1
summarizes some properties of the data we have described. Note
that they are typical properties and individual examples for real-world scenarios,
therefore may differ depending on the application. As we see, the three sources
of data differ in public availability, volume of data usually generated per user,
