Geo-processing: variable buffers to clip data

© A.Krisciunas

^[1] A more sophisticated geo-processing tool would take into account the gravity model theory of business geographic activity. Simply put, the attraction -- and, thus, the volume of interaction -- between two points of human activity on Earth increases by the product of their "size", but decreases by the square of the distance between them. That is :

where: "A" and "B" represent the size of the two places, measured in any number of ways. For example, "A" and "B" may represent the total floor area of two office towers in a downtown area, while the distance may be measured in walking time.

The model, in practise, is considerably more complex (a constant is often applied to account for maximum or minimum thresholds, "intervening opportunities" -- ie: "C", "D", etc. -- must be accounted for, and the exponent of distance is often not "2" but, rather, calculated on the basis of empirical studies of actual interaction), but in its simplest form it still represents a powerful geo-processing tool. At a basic level, variations of it allow the user to determine the area of influence of different types of real estate, particularly in the service sector. The most important factor at play is the concept of distance-decay; that the influence of a site upon its surroundings falls off rapidly as one moves further away from it, regardless of its size.

In the following example, several potential tourism development sites in Nova Scotia are examined.

Challenge: what would be the effect upon local unemployment rates if a selection of tourism sites were developed?

The complication is that these sites are classed into "high", "medium" and "low" levels of potential tourist activity (measured in "peak months of activity"), with concomitant implications for job creation. A single value for the radius of influence, then, is insufficient; we have to account for the differences between the sites.

As with the Winnipeg International Airport example, MapInfo's geo-processing tools will be deployed, where geographic objects (vectors in the airport example, points in this one) in one layer are used to extract geographic objects in a different layer. That is, "buffers" are created in MapInfo, a “geographic query” is then made to select those EAs that “intersect” this buffer, and statistics are calculated for the selected EAs. However, in this example, the buffer radius is calculated as a function of an attribute of each of the points representing their "size", to try and represent the influence of distance-decay.

Again, the data are extracted from Statistics Canada enumeration area (EA) data, contained within a polygon GIS. The 7 sites are geo-coded as points in another layer in the same GIS. The geo-processing steps are similar to those used for the Winnipeg International Airport example, except that:

radii of the buffers will be based on the importance of each site, and
the resulting buffers will then be used as "cookie-cutters" to extract data from the EAs, proportional to the areas of resulting new EA polygons created as the cut-outs by the buffers.

Voila; (click each image to proceed to the next, click logo at top to return to "GIS Case-Studies" page):

^[1]NOTE: If not done so already, the reader is advised to peruse "Sub-Case 2" of the Winnipeg International Airport zoning example, concerning the use of buffers for geographic querying, before going through this case-study.