[Openspace] Order of observations

[Julia Koschinsky]

Davin,

I looked at your shape file's weights matrix in GeoDa's weights matrix
histogram: The number of neighbors is wrong in many cases, which usually
points to the need to clean the shape file to correct for digitizing
problems or problems related to file conversions. We'll try to clean it and
I'll return it with the correct weights.

Julia 


******************************************
Davin Parrott davin at odunsiyusuf.fsnet.co.uk 
Sat Sep 24 06:22:44 CDT 2005 
---------------------------------------------------------------------------

Hello. I am trying to run regressions using data for 72 wards within a 
county. The wards map was imported into GeoDa from a shapefile created in 
Arcview with the wards given consecutive numbers to provide a unique ID. 
When the table is opened in GeoDa, the 72nd ward is the first ward (the rest

then being the 1st to the 71st consecutively). When the contiguity wards 
matrix created in GeoDa is exported (which starts with the 1st ward) and 
used in STATA (which has the 1st ward as the first observation), I am 
getting significantly different levels of spatial autocorrelation (which is 
due to the different ordering of the observations). Is there a possiility 
that GeoDa is re-ordering the observations when it imports the shapefile 
from Arcview? Thanks, Davin Parrott.

******************************************
Julia Koschinsky
Spatial Analysis Laboratory
University of Illinois at Urbana-Champaign
 
http://sal.uiuc.edu 
http://geoda.uiuc.edu