Choose one local origin
Measure every obstacle and observation position from the same reference point. X increases east and Y north, in meters. An obstacle’s X and Y identify its southwest corner; width extends east and depth north. Rotation is not modeled, so approximate a rotated object carefully or use multiple boxes.
Place observation points outside objects
A point inside a solid building has no meaningful outdoor direct-sun result in this model. Put a window test point just outside its façade and use its actual height. The ray test includes height, while the diagram shows shadows projected onto the ground; these are related but not identical views.
Keep the model scope visible
Trees are not opaque boxes throughout the year, and terrain is not always flat. The model cannot represent translucent foliage, glazing or diffuse sky illumination. Use the result to compare a simple geometric arrangement, then check important conclusions against measurements or a more complete professional model.
Decode a complete obstacle row
The example Building | -5 | -6 | 10 | 5 | 6 starts at X = −5 m and Y = −6 m, extends 10 m east and 5 m north, and rises 6 m from the ground. Its footprint therefore spans X from −5 to 5 and Y from −6 to −1. A point at (0,0,0.5) lies just north of it and half a meter above ground.
Enter observation rows as name | east X | north Y | height. Use a consistent origin for every row. Up to 20 obstacle boxes and 10 observation points can be compared. A point inside a box is rejected instead of producing a misleading outdoor sun estimate.
Separate the plan view from a raised observation
The ground-shadow polygon describes where the object blocks sun at ground level. A raised window or balcony point can have a different result, because its ray begins above the ground. Use the point’s actual height and the daily classifications for that question. All boxes begin at ground level and stay axis-aligned; an overhang with empty space beneath it cannot be modeled faithfully as a solid box extending down to the ground.