It's not just about the visuals - its about how the data can be sliced and diced. How your pull requests behave is a very different phenomenon than from how your main branch is behaving. You need to be able to tease this apart easily.