[GRASS-user] v.centerpoint: area geometric median calculations

[Markus Metz]

On Tue, Dec 30, 2014 at 12:03 PM, RichardCooper <richtcooper at hotmail.com> wrote:
> Hi,
>
> I'm using the v.centerpoint addon to determine the* centre point of an area
> (a country vector layer in degrees lat/long)*, and understand from the
> manual that the geometric median will give a more robust answer than the
> center of gravity with regard to possible outliers.
>
> My query: for determining the centre of an area, is either of the median
> calculations (minimum distance to area triangulation or boundary segments)
> preferable? Is one more robust than the other, or is one method more widely
> used?

The answer depends a bit on your requirements. The center of gravity
is the most commonly used center point. Otherwise, I would prefer the
minimized distance to area triangulations over the minimized distance
to boundary segments, but that is probably a matter of taste. If in
doubt, you can calculate both, compare them, and pick the one that
better fits your needs. Both (minimized distance to area
triangulations and minimized distance to boundary segments) should be
equally robust. I am pretty sure that the minimized distance to area
triangulations is not widely available, thus rarely used.

Markus M