By coincidence, I've been collecting just such a data set since Saturday afternoon. I suspect that in the case of the petition I'm interested in, the fact that

By coincidence, I've been collecting just such a data set since Saturday afternoon. I suspect that in the case of the petition I'm interested in, the fact that it had almost no "earned media" before crossing the threshold is why the signing rate appeared to speed up once it had crossed the threshold.