Hacker News Reader

I can accept that in an infographic meant for a wide audience very strong correlation may imply causation. The aim of the infographic is, indeed, praiseworthy.

However the missing y axis for the second factor really upsets me! What's the normalization? How has data been rescaled? That can affect __a lot__ how visually correlated two datasets appear.

View on HN

The problem you're thinking of is when you display two different kinds of data on the same axis and so there's two degrees of freedom in how you choose their relative scaling and offset. This makes it very easy to show "obvious" relationships. But the data here all has a very natural sense in which they share the same axis: they're all temperatures of the world, or simulated values thereof. So the choice in scaling and offset makes little to no difference in how the data is perceived.

Reply on news.ycombinator.com