Good visual designs help you think about causes

Your and your client’s preferred model will usually fit the data.

But there are always other possibilities, especially when you are willing to get imaginative about the explanations. What about other models?

You need to address alternative causal models or explanations as you describe your data visualization. Doing so is a real mark of integrity: it shows your client that you’re not just showing the version of the story that you like best: you’re thinking through possible failure points in your theories.