|Surveying the Territory
Aside from some
hype splatter now and then, the spatial community is benefiting from the renewed
efforts of the spatial data vendors to corner these emerging LBS and real-time
navigation markets. The challenge is keeping up with the various details of each
offering. That said, here are a few tips to keep in mind when evaluating the data
of any big spatial data vendor.
Coverage.In the old days, accuracy and attributes were the name of the
game; they're still important, but also check for such details as coverage outside
the United States, consistency of the data schema, data format options, and
navigational attribute detail - inclusion of information about relative road
height, speed, time-of-day restrictions, and similar rules. In some cases,
what used to be a special option has now become standard. Less common to typical
spatial applications, but perhaps the most interesting options, are the addition
of real-time travel data feeds, voice-enabled datasets, and driver assistance
A dataset's coverage is its geographic extent. At its inception, for example,
GBF/DIME had irregular coverage of the United States, with more detail at selected
urban centers. In areas of recent data-conversion efforts, such as less wealthy
European countries such as Portugal, or in rapidly developing countries like
Brazil or Singapore, a dataset may have the same structure as its U.S. equivalent,
but may contain empty values where the data do not yet exist. Coverage density
also varies. Navtech's European data, for example, include what they call "detailed
city coverage" versus "inter-town coverage" (see Figure 2 and 3). Detailed city
coverage supports door-to-door directions to a specific address, whereas inter-town
coverage connects main road networks for cross-country routing. Check your area
specifically before buying.
Figure 2. Navtech's "detailed city coverage" level provides as many as 150 attributes
per road segment, including addresses, turn restrictions, primary and alternative
street names, physical dividers, one-way streets, and vehicle-type restrictions.
Navtech's "inter-town coverage" provides the main road network in an area,
with select local travel information.
In addition to the complete U.S. coverage offered by all three vendors, there are
variations in international coverage from vendor to vendor. Among their own or
their partners' international datasets, GDT lists Canada, Argentina, and Brazil.
Navtech lists Western Europe, Kuwait, South Africa, and the United Arab Emirates.
And Tele Atlas lists Western Europe, Australia, and Singapore.
Also, the United States is not always the most detailed dataset. For instance, the
Tele Atlas data format is the same in Europe and the United States, with all the same
shapefiles and attribute columns for each. Only in Europe, however, does the model
capture the relationship between public squares and their entry points. Americans
don't have nearly as many pigeon-filled piazzas as the Italians, but if we ever do,
the fields are there, waiting to be populated. This consistency of data schema makes
it easier to write applications for multiple countries, but requires care by the data
administrator in understanding which datasets are valid in each country. In my own
experiences with international spatial datasets, this structural consistency has
proved to be a blessing.
Storage. Another emerging trend is the migration from file systems to
databases for storage of spatial datasets. The ability to store a formerly tiled
collection of files as a single seamless table in a database makes this shift
attractive to managers of large geographies.
Format. Format also can influence the purchasing decision - different vendors
supply different output formats, some ready for immediate insertion to a particular
database. For instance, Navtech supplies an Oracle (
) 8i format and Tele Atlas offers an automated shapefile loading script for IBM
Informix (www.ibm.com) databases.
Before being dazzled by the navigational attributes of any vendor, be sure that
your application software is able to ingest and process that particular dataset's
format. The more complex the model of turn restrictions, overpasses, and time-based
route behavior, the more sophisticated (and costly) the software needed to make sense
of that data. Vendors like Navtech and Tele Atlas recognize this, of course, and
provide their own or third-party tools designed specifically to perform routing
calculations on their formats.
In real-time. After the routine checks are done, don't forget to ring the
bells and blow the whistles. Spatio-temporal data processing is on the rise (see
"Net Results," November 2001) and Tele Atlas aims to lead the charge. Their strategy,
with a product called the Real-Time Traffic Report supplies dynamic content for the
Web or wireless applications by joining text descriptions - those same reports you
hear on the radio each morning - with their actual locations on the map. By
predefining junctions or areas where accidents are most likely to recur and
associating them with an ID value, Tele Atlas can very quickly augment the text
feeds from various traffic-watching organizations with the proper locational
attribute, including an estimate of how long the event will continue. Tele Atlas
plans to expand the service to include real-time weather and traffic speed data
Speak to me, Scotty. User-friendly voice guidance is well-established on
Star Trek, but Navtech may bring it to 21st-century Earth. By recording hundreds
of commands in English, German, French, Italian, Dutch, Spanish, and Japanese,
and equipping their dataset with elements they call "phonemes," Navtech's data
will support navigation systems that listen for verbal commands from a driver
and return verbal replies from the system. Navtech hopes that by 2005, a pure
voice communication system will eliminate all physical contact between human
and machine. Just talk (but don't mistakenly say the word "reverse" while zooming
down the freeway!).
Telematics. Navtech is also busy prototyping a driver assistance system.
Its Web site describes the system as enabling:
Intelligent headlights, which automatically aim to provide optimal
illumination around curves. Adaptive cruise control, which maintains a safe interval
from slower vehicles ahead while maintaining the functions and utility of traditional
cruise control. Smart throttles, which anticipate road conditions (curves and grades)
to save fuel and reduce emissions. Front, side, and rear collision and obstacle
warning and avoidance systems. Warning systems to alert drivers (or eventually even
automatically slow vehicles) when a curve is being approached too fast, when the car
appears to be drifting out of its lane or drifting off the road.
Again, the effort is collaborative, involving programs such as the Intelligent
Vehicle Initiative (IVI) in the United States; the IN-ARTE in Europe; Nextmap,
AGORA, PROBE-IT and research projects with automotive companies.
Telematics and navigational warning systems are not limited to automobiles. At
the ESRI EMEA Conference in Portugal in October 2001, a Saudi Arabian delegate
informed me of his plans to track the course of oil tankers in an effort to detect
and prevent shipwrecks. Some European countries are developing uniform policies to
integrate telematics applications for monitoring water and air pollution, for
medical informatics and health care, and for distance learning into government,
business, and education.
The work at hand. Maybe 2002 will be a brave new year after all. Even the
fear that naturally arises from terrorism and economic decline cannot preempt the
deep satisfaction of a job well done. I don't know if I'd want my car's headlights
to develop a life of their own, especially for $399 a year, but a spatial solution
that reduces emissions is hard to ignore. So keep an eye out for spatio-temporal
data service partnerships and increasingly sophisticated spatial data offerings
in the year ahead. Whether your spatial projects require reference, inventory,
or modeling components, I hope a thorough understanding of today's commercial
datasets helps you select the right basemap data and forge ahead with your meaningful