Understanding your audience
Your audience is one of the most critical things to consider when beginning a new data visualization project. This has two parts: the first is from an editorial perspective (what is the audience's background knowledge of the topic at hand? What types of charts will the audience be able to recognize and properly read? How do these charts work within the broader contexts of this story and other work published?), while the second is technological (what platforms and devices will be used to consume this content?).
It's really important to tentatively sketch out any bespoke data visualization before you start writing code, and this can take many forms. On the one hand, it never hurts to figure out the rough shape of your data before committing to a particular visualization style. Frequently, I've been asked for pie charts with a few small outlier values highlighted, which totally doesn't work (the rest of the chart dwarfs the outliers). You don't necessarily need to...