Pitney Bowes MapInfo Vertical Mapper - Manual (Page 58)

Understanding the Location Profiler Model
56
Vertical Mapper 3.7
This illustration shows the geographic profile of the same table of locations calculated
using only one percent of the total points. The highlighted local concentrations of
points give an effective visual representation of variations in point density across the
map area.
In addition to defining a search area that controls the number of sites used in the distance averaging
calculation, you can also define an exclusion radius that is used to create a circular area
immediately surrounding each grid cell within which site locations are excluded from the calculation.
Using Point Weighting
The Location Profiler allows the use of weighting values attached to points. If all data points are
considered to be of equal importance, then their weighting values will all be equal (typically one). If
some data points are considered to be more important, or perhaps more reliable, correspondingly
higher values may be attributed to them. In real terms, the weighting factor may represent a relative
confidence value assigned to each point if, for example, each site represented a human observation.
The factor may also be a measure of frequency if, for example, each site represented a freight
company’s customer and showed the number of deliveries per month. In each case, the distance
calculation is treated as a multiple of the weighting factor. For example, if the weight represented
frequency of deliveries, then each distance calculation would be handled as a multiple measurement
based upon the frequency value.
Using Distance Decay Functions
A distance decay function can be applied to all points in the location table assuming that a weighting
factor has also been assigned. Distance decay can be used in the Location Profiler model if it is
assumed that the data points close to the grid point are more relevant than those lying farther from
the grid point. In other words, you must decide if it is reasonable to expect that, as the distance to
the data points approaches the outer edge of the search radius, the contribution of these values
approaches zero. A decay value of one is suggested for data points close to the grid node, and it
falls to zero for data points closer to the edge of the search radius. A distribution of this type typically
has an inflection point somewhere along its length. An inflection point is the point along the
distribution where the slope either stops increasing in value or stops decreasing in value. For typical