If you chart response count (either as numbers or probably more likely as a percentage of responses) over time, you can add a filter to the count metric. So if you have one count metric with a filter that only captures Promoters and another count metric with a filter that only captures Detractors, I think that's what you want.