When we want to find the area of a two dimensional figure and there are more than one type of shapes that can be used to find the whole figure, we find the areas of the shapes comprising the figure one step at a time. Then finally adding the areas of the shapes together to find area of the whole figure. But do we round twice or just at the end of the calculation? And does it always have to be rounded to the least number of significant figures. What does this achieve exactly? Please show an example and show your reasoning.

-Thanks