Geography Reference
In-Depth Information
months before for planning, analysis, and testing. And, by extension, $10,000 should have been spent
before that to determine how to spend the $100,000. The 10 percent planning rule suggests, then, that a
small but substantial and representative amount of data be collected, encoded, and validated before the
major data collection effort gets under way.
The process of data collection must be carefully planned and executed. It is possible to spend a
lot of money at it and wind up with nothing very useful. Among the points to consider are the
following:
(1) Some work may well be contracted out. Certainly all understandings with the contracting firm
must be written down. As important is that such understandings are completely comprehended
by both parties. (It's not hard to mess this up. (Famous example: The Mars Climate Orbiter
crashed into the planet instead of cruising around it because NASA and a contractor miscom-
municated—using different units (English instead of metric).)
(2) There aren't very many firms that do good intermediate- and high-altitude orthophoto work.
Those that exist may be scheduled for months or years in advance.
(3) Rigid timetables for collection of data about the environment cannot be followed. Clouds form,
trees get leaves, airplanes malfunction, the ground gets wet. Timetables must be based on
probabilities.
(c) Consider datasets that are collected in an ongoing fashion by satellite or aircraft. Datasets showing
features in color at resolutions on the order of a meter are now available—available, but not cheap.
These datasets primarily depict land cover (from which, in many instances, land use may be deter-
mined). The use of data collected in this way has a number of advantages. Among them: (a) the data
may be obtained in already digitized form, (b) updating takes place on a periodic basis, and
(c) sophisticated computer software is available to manipulate these data.
5. Put the data acquired into proper form for inclusion in the GIS database.
Basically, the process of encoding the data means transforming it from the basic form in which it is
collected (or acquired if already collected data are used) into the symbolic or graphic form required
by the GIS. The process depends on the types of data, the precisions required, the equipment avail-
able, the scheme used to represent the data in computer memory (the storage paradigm), and other
factors, discussed in Chapter 4.
6. Check the accuracy of each step of the collecting or reformatting process; also check the first form
of the data against the form in the GIS.
Two elements must be constantly monitored: (a) the process of collecting and reformatting the data, and
(b) the quality of the data sets themselves. Perhaps the most important statement that can be made about
this “checking” process is that it be accomplished by someone other than the person or group doing the
collection.
Such independent checking has many virtues: it provides for a more objective view by those doing the
checking; it ensures that the checking activity is a project in itself and not just an adjunct to the data
collection effort; and it reduces the temptation to use the same techniques to check the data as are used to
develop the data.
There must be more to the checking process than simply ascertaining and reporting error rates. An
understanding of why errors occur must be developed. If the encoded value for evaluation at a certain
Search WWH ::
Custom Search