The trouble with D3 is the banner right at the top of the homepage, full of beautiful visuals. Beauty is great, but clarity has to come first. And clarity is

The trouble with D3 is the banner right at the top of the homepage, full of beautiful visuals. Beauty is great, but clarity has to come first. And clarity is partly driven by what the audience is familiar with. For most audiences, scatter plot, line plot, bar chart, and maybe parallel coordinate plot. Possibly a map as well, although plots on top of maps are famously easy to misinterpret if you don't know what you're doing when you make them. Engineering disciplines have their own specialized visualizations like Smith Charts and carpet plots. The complexity of designing a plot comes with how you compose together the basic elements. The consumer of the plot will want insets or annotations, two plots overlaid or juxtaposed. Error bars. It's very easy to get overwhelmed if your starting point is trying to design beautiful visuals. D3 will let you build all the needed complexity, but if you start by copying a beautiful plot, you'll soon find yourself overwhelmed, unable to deal with beauty and complexity at the same time, struggling to reconcile new functionality with the initial design. My advice is to make the simplest possible plot of your data (one of the basic types, monochrome if you can get away with it), and figure out whether you need to add anything once you and your customer have tried to work with it.