1

I have data for some countries but not for others and would like to predict the missing values from neighboring countries weighted by their area, distance etc.

To do so, I am using the geostatistical wizard to interpolate information between polygons. Since I have not even succeeded in implementing areal interpolation using the ArcGIS interface, I did not dare to use ArcPy which is why I cannot provide you with any code. Here is what I did:

I downloaded the world country geodatabase available from ESRI (https://www.arcgis.com/home/item.html?id=3864c63872d84aec91933618e3815dd2), merged it with my data and then used the tool Areal Interpolation from the geostatistical wizard.

At step 2, I received the error 40066: cannot distinguish between polygons which presumably occurs because some polygons (islands) are much smaller than others. As suggested in the ESRI help, I used the tool Delete Identical to identify and delete those polygons that ArcGIS could not distinguish between. Apparently, this was successful but when I tried to continue with the Areal Interpolation I got the same error 40066 again.

Did anyone have similar problems?

  • 1
    What format was your original data in, and how exactly did you merge it with the country polygons? Do you have one value per country, or are there many values? Also, did you try some of the other strategies from ESRI help (Polygons of vastly different sizes)? They suggest adjusting the lattice size or not including small polygons in the analysis: desktop.arcgis.com/en/arcmap/latest/extensions/…. – Priscilla Dec 6 '16 at 16:36
  • Thank you very much. I had tried various approaches (using shp or using geodatabase) as well as some of the suggestions from the ESRI help (eg. adjusting lattice size). What did the trick for me was your last suggestion, that is simply not including smaller polygons in the analysis. Thanks a lot. If you write down your comment as an answer, I will mark it as accepted. – eigenvector Dec 8 '16 at 13:04
1

The ESRI help on the Areal Interpolation Tool suggests three potential solutions (See subhead 'Polygons of vastly different sizes' under Limitations).

  • Use Find Identical and Delete Identical tools (which you already tried).
  • Lower the lattice spacing
  • Don't include very small polygons in the analysis

You'll probably have to trial and error with the lattice options or the level of polygon size to filter out.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.