How is proximity filtering performed?
Proximity filtering results are calculated by running spatial queries based on a selected distance of a point or points. Two types of distances are currently used within SpatialKey: standard units of measure such as feet, miles, meters, kilometers, etc. (converted to meters internally) and degrees of latitude and longitude (the native way the dataset is stored within the SpatialKey spatial database).
As of 11/1/2009 the results of these queries may show as an elongated ellipse (if degrees of latitude and longitude are used) or an actual true circle (with a radius based on the user’s entered value and unit of measure) depending where on the globe both the dataset(s) being filtered and the dataset doing the filtering are located.
Datasets within the United States (and most of North America)
For datasets that are completely contained within the United States a direct conversion of the user entered distance can be used to create a true circle result. Internally we are able to convert the stored data from latitude and longitude degrees based measurement into an accurate meters based representation for the filtering (see figure 1.1 for an example).
Fig. 1.1 Proximity Filtering with 10 km (not degrees)
For this example we are filtering from the city center of Sacramento, CA and have set a distance of 10 kilometers. The outer ellipse shows the boundary that would be created from this search if degrees of latitude and longitude were used while the inner circle is a true representation of 10 kilometers as a radius from the city center. In this case, the results are clearly within the 10 km circle as the conversion process was utilized. Also note that the difference between the ellipse and the true circle become greater with longer distances entered by the end user.
The conversion that we use in this process only works for points that exist within the United States (and some other parts of North America, for more information and to see the extent of this “projection” click here). The SpatialKey client will automatically detect if this conversion is available based on your data and use this version if possible.
Datasets outside of the United States
For all other data outside the United States the degrees based calculations will be used. This will result in the elongated ellipse based results as shown in figure 1.2. The shape and accuracy of the ellipse will depend on the location of the points on the globe as latitude is always a fixed distance at 60 nautical miles per degree, but one degree of longitude can vary from about 60 nautical miles at the equator to zero at the poles.
Fig. 1.2 Proximity Filtering with 10 km converted to degrees
Here we see that additional points are now included in the results as the degrees of latitude and longitude version of filtering was applied.
We will continue to improve the proximity filtering within SpatialKey by adding in additional conversion to other parts of the world in future versions. When these conversions become available they will be automatically used (just as the United States version is today) for the filtering queries.
Follow Us: