GeoMedia GIS Blog

"A picture is only worth a thousand words. A map may be worth a thousand numbers. But a GIS is worth a thousand tables."

Nearest Neighbor Aggregation in GeoMedia 6.1

Posted by jeffhobbs on June 27, 2007

Started playing with the two new aggregation operators (pipes) in GeoMedia 6.1 today. There appear to be a few bugs, but overall it looks very promising.

For my testing I used the US Sample data set and aggregated the cities and the highways. My summary feature was the cities feature class while my detail feature was the interstates feature class.

Aggregation Dialog


The first operator is labeled "are nearest to". From what I can tell (no documentation yet), this operator will aggregate ONLY the first summary graphic to its respective detail graphic. In other words, if I have 20 cities that are near one highway segment (not the overall highway, but the individual segment), then only one city will receive the respective highway number from the highway segment. Looking at the screen shot below, you’ll notice that each highway number (the second column that’s highlighted) has a respective unique identifier (third column), however no unique identifier is repeated for the same highway number.

Now, when comparing this to the second operator, "accumulate nearest", the opposite is true. With this operator EVERY city has a respective nearest highway. As a result, you DO get repeating highway numbers (second column) for each unique identifier.

The Value

Now the value…that can be a little harder to understand. I personally have never had a need (that I can think of) for an operator like "are nearest to". Anybody have any ideas?

However with the "accumulate nearest", this to me makes a lot more sense. Often times I’ll want to bulk update a set of geometries in feature class B with information from feature class A. For instance, you might want to update your fire hydrants with the nearest one or two street names. In the past, if you’re an Oracle Spatial kinda user, this can be done at the SQL level. However for those that would rather do it in the client GIS, this makes it a lot easier.

Until tomorrow…..

6 Responses to “Nearest Neighbor Aggregation in GeoMedia 6.1”

  1. Shelley said

    My colleagues are currently working on an algorithm to calculate nearest address point to a point feature, for example fire hydrants in the road or any other point feature that does not usually sit on Land Parcels.

    So that you can give a direction to someone on where to go to find something that does not have an address.

    “The Hydrant is near 10 Smith Street.”
    “The accident happened outside 15 John Street.”
    “The body is buried near 20 Jones Street”


  2. Jeff Hobbs said

    Are they doing this in GeoMedia Objects? This almost sounds like what could be a customization built off of 6.1.

    Very interesting!

  3. Shelley said

    They are building a series of queries, including using Grid analysis like Thiessen polygons and the Buffer with cost, to create a polygon for each address point.

    When the logic of the queries is finalised, we will produce this manually on a regular basis and derive the relationship with with other features automatically overnight.

    Eventually we will automate the producing the polygons.

  4. Jeff Hobbs said

    Very interesting. This is being done in GeoMedia Objects (as opposed to something like Oracle Spatial), correct?

  5. […] video demonstrating the new nearest neighbor aggregation in GeoMedia 6.1. If you recall, I had an entry on this towards then end of […]

  6. Shelley said

    Yes :o)

Sorry, the comment form is closed at this time.

%d bloggers like this: