# Buffering a PointGeometry in arcpy only returns a line?

I am trying to buffer a `PointGeometry` object in arcpy to a `Polygon` geometry object with ArcGis 10.2.

The `PointGeometry` object is in lat/lon WGS84 coordinates and as far as I understand, I can only use `PointGeometry.buffer` by specifying the buffer in the same coordinates. To specify the buffer in metres, I would need to use: `arcpy.Buffer_analysis`.

So I setup an empty geometry to do the buffer in like so:

``````spatialRef = u"GEOGCS['GCS_WGS_1984',DATUM['D_WGS_1984',SPHEROID['WGS_1984',6378137.0,298.257223563]],\
PRIMEM['Greenwich',0.0],UNIT['Degree',0.0174532925199433]];\
-450359962737.05 -450359962737.05 10000;-100000 10000;-100000 10000;0.001;0.001;0.001;IsHighPrecision"

pointGeom = arcpy.PointGeometry(arcpy.Point(-27.607361, 152.768311), spatialRef)
temPol = arcpy.Geometry()

pol = arcpy.Buffer_analysis(pointGeom, temPol, "50 meters")[0]
``````

The resulting `Polygon` object however, is a long thin line, rather than the expected circular polygon.

Any ideas what I am doing wrong?

• What version of ArcMap are you using, and where/how are you defining `spatialref`?
– Paul
Commented Aug 5, 2014 at 14:36
• @Paul : updated the question with the answers Commented Aug 5, 2014 at 14:38
• You have your X,Y values flipped -- 152.8 degrees north is messing up the buffer. Commented Aug 5, 2014 at 14:49
• Thanks. So `Point` expects lon, lat rather than lat, lon? I'm now getting polygons. But they are multisided, irregular polygons....I would've expected circles for a single point? Commented Aug 5, 2014 at 15:18
• `Point` expects (x,y). Circles aren't circular in decimal degrees (which is a different question). Commented Aug 5, 2014 at 15:27

1. The `Point` constructor takes arguments of X and Y, while you're providing Y (latitude) and X (longitude), resulting in an object above the north pole
2. The XY precision you've defined in the `spatialRef` object is four decimal places (1/10000th of a degree), which is exceedingly coarse for a circle with a 50 meter radius -- an XY scale of 10,000 corresponds to roughly 11 meters (at the equator)