Geographic and Attribute Data: Selection, Input, and Editing - Introducing Geographic Information Systems with ArcGIS - page 287

Assuming a set of already collected data has been found that approximately meets the specifications, the
data should be carefully analyzed according to the number of characteristics:

Can the data sets be used directly or must they be manipulated before they can be used in analysis?

❏

Are the data the most specific and detailed available? When accuracy is critical, and you have to
resort to digitizing maps, can you find a version printed on nonshrinking, nonstretching Mylar?

❏

Is the resolution of the data sufficient to fill information needs?

❏

Are data mapped at an appropriate scale for the resolution required?

❏

In what geographic coordinate system are the data recorded? What complications will occur in con-
verting the data if conversion becomes necessary? Will accuracy or resolution be lost in the process?

❏

What process was used to collect the data? Do statements about the precision and accuracy of the
data accompany the dataset? Did the creators of the data sets seem to take the idea of metadata
seriously?

❏

Are the data uniform? Is the medium on which they are recorded also uniform and free from the
kind of distortion found in some un-rectified aerial photos?

❏

Are the datasets truly available? Who owns them? Are they in the public domain? Can “originals” be
obtained, or only copies; what information is lost in the copying process?

❏

What process was used to collect the data? Are they subject to provisions of confidentiality? Are they
classified by the military?

❏

How much time will be required to obtain the datasets? How much time must be allowed to refor-
mat or encode them?

❏

Will information updates be available from the same source? If not, will new updates, possibly from
different processes, mesh with existing data?

❏

When all considerations are combined, what will the data cost?

❏

(b) If you can't find data sets that meet your specifications—or even if you can—you may embark on
a data collection effort. In many ways, if data sets that approximately meet your specifications are
available, the issue of whether to use them or collect your own is much like the issue of whether to
buy a used or new car. There are advantages, disadvantages, and uncertainties associated with both
courses of action. The decision can become very much the classic avoidance-avoidance conflict that
college sophomores learn about in psychology courses: the more you look at the other people's data
for your requirements, the more you want to collect your own; the more you examine what you have
to go through to collect your own data, the more attractive the existing data sets seem.

If you decide to collect data anew, many of the concerns for characteristics still apply, but the question
changes, from “Do these data have the properties I want?” to “How do I construct a process to produce
the data and characteristics I want?”

Probably the best advice to anyone planning a large data collection effort is to start slowly . . . and
carefully. In fact, with all operations involving a data-handling program, one should probably use a
“10 percent planning rule.” This rule says that if x dollars are to be spent over a period of time, then10
percent of x dollars should be spent over a previous period on the same subject. For example, if 1 million
dollars is to be spent on GIS data development in a year period, $100,000 should have been spent in the