The colored signal level vs az/el plot  follows that convention...  outer edge is the horizon, center point is zenith (90 degrees above the horizon).  Color represents the average signal level seen at a given point in the sky.

The tricky bit is interpolating signal levels between logged points (there is a display options for showing the raw signal level data).  Heather interpolates between adjacent azimuth points at each elevation angle.  The interpolation extends for a maximum of 22.5 degrees azimuth from each point that had signals.  Handling things like the "hole" at northern azimuth angles (for the northern hemisphere) complicates things (once you get far enough north, the horseshoe shaped signal hole becomes a closed circle).  Heather builds a "clipping" structure based on the min and max elevation angles that has signals for each azimuth angle.

I'd like to try something like a Voronoi tessellation,  but that gets rather nasty to implement...

radius = 90-elevation angle (so horizon is outer border), angle is
azimuth looking down on antenna, and color be power.

Mark wrote:

I always wondered about the "filled in" (interpolated) plot.  To my
observation, the orbital tracks of the satellites do not wander anywhere
near +/- 22.5 degrees.  It is unclear to me that interpolating the data
to areas of the sky where satellites never appear has any utility.

I suppose it makes trends somewhat easier to spot -- but that could also
be done just by making the dots on the raw chart a bit larger, without
suggesting a continuity of reception-space that doesn't exist.

Or am I missing something?

Charles

rather nasty to implement..

There's a boost library...