There is a file called merged_space.map in the numeric directory of the radiance measure's run dir; this is a concatenated list of all the illuminance maps found in the OSM, translated to Radiance rcalc inputs. These inputs take the form x y z vx vy vz, e.g.:

0.305 0.305 0.762 0.000 0.000 1.000

In the example above, there is an illuminance map point defined at 0.305 meters in the x and y from model origin (world coordinates), 0.762 meters off the floor, and the calc point is looking 'up' (view vector is 0 0 1).

merged_space.map is the input to Radiance, so the output file (output/merged_space.ill) relates to this input file. IOW, each row in the .ill file after the header is an 8,760 csv list of daylight illuminance for each point found in numeric/merged_space.map, in the same order. You could visualize the data spatially with any number of tools (R and friends, D3 or whatever Javascript flavor of the week is). We have done this internally with R for checking the integrity of the output of the Radiance measure, simply mapping the x y z from merged_space.map to the value(s) found in the corresponding rows in merged_space.ill. Make sense?

By the way, if you have shading controls on your model and save the individual window group data you can visualize each window group's contribution to the whole building for each shade condition. This is awesome and cool. Again, a visualization interface would be somewhat easy to develop and would pay huge dividends. This is an exercise for the user, as my friend @GregWard has said before. =)