I've responded to your comment there. Obviously, the data is open to lots of interpretations. I tried to use models that seem reasonable and reflect what we kno

I've responded to your comment there. Obviously, the data is open to lots of interpretations. I tried to use models that seem reasonable and reflect what we know about internet behaviour - particularly the Pareto distribution of most activities.