Information Technology Reference
In-Depth Information
Difficulties in bringing theory and application researchers together prevail. Some
application areas (for example, movement ecology, crowd management) maintain
and further develop CMA research fields with ample momentum but little links to
GIScience and computer science.
Multi-sensor systems (speed, acceleration, physical properties) replace pure loca-
tion tracking systems.
After a cooler phase, privacy recovers as a key CMA topic.
After the golden first decade of CMA, progress seems to be slowing down as the
low hanging fruit are gone and the hard problems prevail. Even though movement
analysis is still prominent on most schedules of relevant conferences, I would argue
that the research field gains breadth rather than depth. There seems to be no bound to
the arrival of new and and fascinating movement data sources, each with a specific
problem producing a unique movement analysis solution. Even though any growth
benefits the field, if the field wants to mature, a set of grand challenges need to be
addressed.
Clearly, aspects of the following relate to old and well known challenges of han-
dling spatio-temporal data, such as scale, uncertainty, or data integration. Other
aspects of CMA pose new and interesting challenges, opening up countless research
avenues for the years to come.
5.1 Coping with Big Movement Data
GIScience and geomatics are currently experiencing an exciting revolution, that also
has implications for CMA. What started off as a data poor and computation poor
discipline, then coped with data rich and computation rich environments for decades,
is now faced with another dramatic shift regarding its data sources: Big data (Graham
and Shelton
2013
; Kitchin
2013
) Sensor networks and mobile GIS, user-generated
content and open data, the web and cloud computing inevitably change how we
capture and manage geoinformation, how we analyze and exploit geoinformation,
and ultimately how we take decisions based on geoinformation.
These new and massive data streams not only challenge CMA in terms of volume,
but also require new ways of integrating heterogeneous multi-source information
(Fig.
5.1
). Such integration could, for example, require inferring behavior from GPS
tracks sampled at one second intervals, combined with accelerometer readings that
come in 6 s long bursts sampled at 20Hz, but only every 10min, and interpolated
and hence uncertain meteorological field data with a spatio-temporal granularity of
daily values per 1km grid cells. Big geodata as a source for CMA means inferring
knowledge and making decisions based on more comprehensive but at the same time
more uncertain, messy and noisymovement data. If and to what degree big data really
poses fundamentally new challenges is still widely discussed. One could argue that
especially GIScience with its long tradition of handling voluminous and messy data
is in an excellent position for contributing to the big data debate.
