Percent changed compared to 2015 would be better, but choosing a graph like this makes no sense if it would have been change yoy and also if that were the case then starting at 0% makes no sense.