This is a good example of why I think any visualization of correlation should include the actual Pearson correlation value. I don't think the correlation here is a strong as the title makes it out to be (wouldn't be suprised if < 0.5), even setting aside concerns about whether it is spurious.