Select to view content in your preferred language

XY to line/distance between matched points

1007
2
01-08-2013 05:24 AM
R_Acron
Emerging Contributor
Hi,

I'm trying to find the distance between matched points, or to somehow compare the results of one geocoding method with another to see how close the results are. I have tried the method using XY to line here:

http://kb.mit.edu/confluence/pages/viewpage.action?pageId=11338190

using both a geographic coordinate system and a projected coordinate system. The unique ID field that I choose is never populated in the resulting attribute table, and one set of x and y fields is "<NULL>" -- which makes me think that I'm doing something wrong, but I have no idea what. Any suggestions would be appreciated.

Thank you,
Robert
Tags (3)
0 Kudos
2 Replies
R_Acron
Emerging Contributor
Thanks Luke.
0 Kudos
R_Acron
Emerging Contributor
Just adding, in case someone else runs into the same problem. I can't pinpoint exactly where I went wrong at this point, but the problem I had seemed to come from not converting between feature classes and/or consolidating my data correctly. I'm pretty new to ArcGIS...

What I ended up doing was creating two shapefiles (each with their own X and Y coordinates), joining them on the unique id, and then exporting the attribute table resulting from the join to a .dbf before using "Run XY to Line", which seemed to produce reasonable results.

I'm planning to look into the Bivariate K function too though. I have only used R slightly + quite a while ago, so it will definitely be a challenge.  

R.

Hi Robert,

I would approach this problem with spatial analysis of point patterns. The first thing that comes to mind is the Bivariate K Function. This would test if the two point datasets are different or not. This link should be a starting point: http://wiki.landscapetoolbox.org/doku.php/spatial_analysis_methods:ripley_s_k_and_pair_correlation_f.... I do not believe that ArcGIS has a tool for Bivariate K Function, but you could use R and the package spatstat to do this analysis. You could also potentially program this in python, but using R I think would be simpler. You also might think about using kernel density estimation as well on both datasets.

I hope this helps,
Cheers,
Luke Kaim


Thank you,
Luke Kaim

Thank you
Luke Kaim (SIE)
Lucas.Kaim@maine.edu
(914)263-7866
"Don�??t complain. Just work harder" (Randy Pausch).
0 Kudos