Perspectives on Space and Time in US and Chinese Science - Space-Time Integration in Geography and GIScience - page 11

ground-based view from the current location of a smart-phone user. Time has new
importance, and is now increasingly integrated into GIS and geographic information
generally. Yet it has proven difficult to escape the power of the map metaphor
(Goodchild 1988 ) and its constraining influence on research and development. In
the case of the automobile it took many years before the metaphor of the horse and
carriage (the horseless carriage ) could be dropped and the full potential of the new
technology realized, and this same pattern has clearly influenced the integration of
time into GIS.
In the US, researchers now have access to a vast flood of geographic information,
a significant part of what is variously referred to as the exaflood or Big Data .
The locations of aircraft, buses, fleet vehicles, and ships are now continuously
monitored and in many cases made available through the Internet. Mobile phones
are continuously tracked when turned on, with positional uncertainties as small as
a few meters. Credit and debit cards produce spatiotemporal records whenever they
are used, though these will typically be available only to banks and law-enforcement
agencies. Imaging sensors on satellites create terabytes of geographic information
daily, as do sensors on aircraft and unmanned drones. Ground-based sensors monitor
traffic conditions at points on road networks, and companies such as Google collect
ground-based images from moving vehicles and even bicycles.
While these new data are abundant, major issues exist in linking and synthesizing
them into the forms needed for research and applications. Traditional research
sources, such as the US Census, are collected subject to rigid procedures and norms,
compiled using replicable processes, and documented with detailed metadata. As a
result it has been possible for vast amounts of research to be conducted on them over
the years. But the new data sources often are incomplete, with no rigorous sampling
design, and little in the way of metadata or quality assurance. Elwood et al. ( 2013 )
have argued that this new world of Big Data will be dominated by the need for
synthesis and post-hoc quality control, and will require the development of a range
of new techniques.
These arguments are especially apposite in the case of geographic information
developed and contributed by citizens, as distinct from qualified experts. So-called
volunteered geographic information (VGI; Goodchild 2007 )or crowd-sourced
geographic information is exemplified by projects such as Open Street Map and
Wikimapia, and now constitutes a significant source of knowledge about the
geographic world that is comparable in its value to that produced by the traditional
mapping agencies. Many companies now rely on VGI to update their digital street
maps, and mapping agencies such as the US Geological Survey make use of citizen
input as well. Goodchild and Li ( 2012 ) have reviewed the available strategies
for quality assurance, and the growing literature on VGI quality has often shown
accuracies at least as good as those of traditional sources, especially with respect to
timeliness.
It was argued earlier that maps traditionally emphasized stable phenomena such
as buildings, streets, and topography over transitory phenomena such as traffic
conditions. Map-making was expensive, requiring the fielding of crews of experts.
But VGI has the potential to provide timely updates and observations of phenomena