Greetings,
I am working on an interesting problem in economics involving ranking a
set of several hundred land parcels according to their potential for
commercial building development. I have collected approximately 30
variables for each parcel, representing a number of economic,
locational, statutory, and descriptive characteristics. My goal is to
produce some sort of statistically-based summary measure, which for
lack of a better term, I will call an "Index". Ideally the Index
should have the property of providing an ordinal ranking of development
potential, based upon an assessment of the set of descriptive factors.
Thus, parcels with the highest rank would offer the most potential (in
the sense of commanding the highest sales price). At the end of the
analysis, some top percentile of land parcels will be identified for
more detailed analysis.
One feature of this problem is that the ranking criteria -- development
potential -- is unobservable. We have a set of attributes hypothesized
to influence the ranking, but we cannot observe the potential itself.
My intuition is that this problem might be amenable to some sort of
factor analysis, but I am not sure. I am hoping that list readers might
offer some suggestions about how this ranking could be constructed.
Thanks,
-Mark
|