Geography Reference
In-Depth Information
4.1 Types of Spatio-Temporal Clusters
In her ''Mapping Time'' monograph, Vasiliev (
1997
) defines five time categories
that are used by geographers: moments (moment in time, single instance, dating of
an event, no duration), duration (intervals, continuance of an event in space),
structured time (sequences, ordering of events, organization of space by time),
time as distance (time as a measurement of distance) and space as clock (space as
a measure of time). These categories define different time interpretations and
representations on the map. We derive our taxonomy using the basic definitions of
these categories where moments, duration and structured time are the most
important for our definition of types of spatio-temporal clusters.
Note: our definition of events is similar to that of Becker et al. (
2009
)—an event
is something that occurs in a certain place at a certain time and characterized by
some photo activity of people. For the more general definitions of events in geo-
spatial domain we refer the reader to Worboys and Hornsby (
2004
), Cole and
Hornsby (
2005
) and Hall and Hornsby (
2005
).
1. Stationary (moments)—the cluster is called stationary when the subject of a
photograph does not change in time. Landmarks like a monument, museum or
airport are good candidates to be found in such a cluster.
2. Reappearing (duration, intervals)—clusters can be reappearing when the pho-
tographic activity in the area increases in one time period and decreases in
another. Two types of reappearing clusters can be expected:
a. Regular—clusters in which some periodic events take place. Such events
attract people at regular periods. For example: new-year fireworks on the
main square of the city or Oktoberfest in Munich.
b. Irregular—clusters in which aperiodic events take place. Such events attract
people at irregular periods. For example: a concert or football game.
3. Occasional (moments)—clusters in which some events happen occasionally,
e.g. a traffic accident.
4. Regular moving events (structured time)—clusters that represent the same
event taking place in different places, e.g. Olympic Games (taking place on a
regular basis in different places) or scientific conferences. This is probably the
most complex type of a cluster since, for each particular place, it is irregular or
occasional, and can only be discovered by inter-place comparisons.
In addition to the types of spatio-temporal clusters, we would like to differ-
entiate between two types of semantics: (1) semantics of places and (2) semantics
of user's behavior. Semantics of places and users are interconnected and one can
enrich the other. Knowing the semantics of a place, we can infer the semantics of
users who take photos in these places. For example, the place that has semantics of
sport (stadium, football) will indicate that people who take photos there like sport.
Likewise, characteristics of a user can have implication on the analysis of the
event, user's behavior or a place. Using the profile of the person, we can interpret
