Today I am cheating a little. In the last post, we created a symmetrical distance matrix using strict distances between 6 cities. I promised that we would create an Inverse Distance matrix, which is used extensively in economic geography. I say that I am cheating because to create the Inverse Distance Matrix, you simply place a 1 over the computed distance:
TRANSFORM 1/Min(int(DistanceEarth([Cities].g, [Cities2].g, “mi”))) SELECT [Cities].name FROM [Cities], [Cities] AS [Cities2] GROUP BY [Cities].[Name] PIVOT [Cities2].[Name]
This creates an inverse distance weight, meaning the closer the cities are to one another, the larger the weight.
Want to create an inverse distance squared matrix? Just add an exponent (^) to distance:
TRANSFORM 1/Min(int(DistanceEarth([Cities].g, [Cities2].g, “mi”))^2) SELECT [Cities].name FROM [Cities], [Cities] AS [Cities2] GROUP BY [Cities].[Name] PIVOT [Cities2].[Name]
the results of this query is:
Comments