This is a really, really bad graph.
1) It shows cumulative events rather than rate. While it is possible to extract rate information (slope), it is mingled with noise and problem #2
2) Destructive events are almost certainly seasonal, This graph is zoomed in too much, making it hard to disentangle seasonal trends from scary trends.
This would be so much better as a table: X in 2019, 2020, 2021 and Y in 2019, 2020, etc....
(post is archived)