[Taxacom] Data quality of aggregated datasets
deepreef at bishopmuseum.org
Tue May 7 12:14:52 CDT 2013
One thing I might not have made clear in my previous post: storing locality
data in the form of a custom bounding box (e.g., where you record two
lat/lon coordinates, and use them as two opposite corners of a rectangle) is
essentially the same as point-radius, so this isn't an issue of circles vs.
squares. The point is that when you capture or interpret data based on a
pre-defined grid (UTM, or some other grid), you lose more information than
you need to. Regardless of whether I had a GPS when I captured a specimen,
or whether I'm trying to georeference a specimen label that says the
specimen was taken "Along a trail 0.25 mile SE of the big tree at the end of
Mulligan's Road", it's better to capture and store the data with as much
accuracy as you can, then filter/convert it as appropriate into an arbitrary
grid as needed for analysis.
More information about the Taxacom