April 26, 2018
- Lines are clearly differentiated
- It is intuitive
What could be improved
- I’ll work on the title
- Remove the legend (test for color blindness)
- Right align the y-axis labels
- Remove vertical grid line
- Add full years
- Add annotations at points where lines intersect
- Research proper x-axis / y-axis ratios
A few fun tips
I use Spectrum to test for colorblindness.
You can see the original colors work well, buuut I decided to go with some different colors anyway. Notice how I said, “I decided”. My color choice is totally subjective. Without colors that more clearly associated with the categories in our dataset, it’s all free-range (pun intended) less those colors that make it hard for others with color vision limitations.
While most colors are free to use, notice what happens when the low-contrast option is selected. Those grid lines become a bit pointless, hey? There’s your justification to remove the background color. Every choice is a design choice in data visualization.
My favorite tool to use top pick pretty colors is ColorZilla. Regardless of the tool you use, ColorZilla makes it super simple to grab those HEX codes your visualization tools needs for custom colors.
- The golden ratio
In order to make my visualization (technically) beautiful the dimensions need to be just perfect 👌 Visua.ly has a good slide deck you can run through to learn more about the golden ratio if you are into that sorta superficial stuff.
No, seriously, the reason I use this ratio in my line charts, besides the beauty, is to avoid skewing the data.
An x-axis that is too long will shrink the slope of your lines making spikes seem less severe. Create an x-axis that isn’t wide enough (too narrow) might make your audience think the rise or fall you are presenting is more drastic than it should seem.
That’s it for this week.
Let me know what you learned from reading this on Twitter @robcrock
Maybe you want to collaborate? I could be into that. Say hi at firstname.lastname@example.org