It's not that obvious with the 2009 Sello shooting being dropped out and the 2011 Utoya shooting being included.

Given the poor coverage it is very difficult to reverse engineer what this graph is supposed to portray without finding the original source.