I'm getting consistent, weird results from using Minimum Bounding Geometry to make circle polygons. I'm not sure if a datum issue, but I'm leaning more towards something wrong with Minimum Bounding Geometry tool. Here's the situation:
Goal is to acquire diameter in meters from minimum bounding polygon circles.
Python script works as follows:
- Import Lat/Lon values with MakeXYEvent Layer (setting GeoCoordSys to D_North_American_1983)
- Copy Features
- Project points (to North America Albers Equal Area Conic)
- Create Min. Bound Polygons around groups of points (ranging from 2 to over 50 points in a group).
Results seem reasonable for most polygons, however there are a good portion that the resulting circles wrap the globe. In several cases the points are clearly grouped in a small area off the coast, and the polygon is covering an enormous area of land, not even overlapping points. This in combination with "inconsistent extent" error makes me think it's a datum/project issue.
Troubleshooting efforts so far include
Points used to create the erroneous min bound circle outputs do not appear to have unusual lat/lon values.
Running Min. Bound Polygon to create a convex hull on points that were consistently creating crazy circles. Convex Hull created seemingly reasonable results.
- Using convex hull polygon as input for min bound circle, and created same enormous circle.
Have you seen this before? Is this a case of choosing the wrong projection/datum? An error in Arc? An error in Min Bounding Geometry tool?
Thanks for input!