Welcome to issue #133 of The Process, the newsletter for FlowingData members where we talk about how the charts get made. I’m Nathan Yau, and this week I’m trying to reduce the number of choices when I work with data.
But first, a quick note for all you ternary plot fans out there. I’m sure there are thousands, nay, millions of you.
Maarten Lambrechts is back with a new tutorial on how to make ternary plots in R, using ggplot2. The plot type is commonly used to compare composition where three parts add up to a whole. Maarten shows you how, plus takes it a step further by showing data over time and annotating for readability. Enjoy.
Back to those choices.
Some folks were lamenting about how data type and format is one of the first things you must specify when using various visualization tools. Import your data, the software tries to guess the data format without contextual information, and then suggests a handful of chart types.
If you only consider data format to choose your chart type, then yeah, you’re probably approaching the problem too mechanically.
It’s a way to narrow down the possibilities.
When you first get a dataset, there are many possibilities and small decisions along the way to get you to a finished chart. It’s an infinite space. You can’t possibly find the best choice out of infinite possibilities, and the challenge can feel overwhelming.
But introduce constraints, and the space of possibilities shrinks, which makes it much easier for you to decide what to make.
I do this every day when I decide what to make for dinner. I open the refrigerator and look at the ingredients available. If I had infinite ingredients that were instantly available, then I’d skip the step and immediately consider what we feel like eating and how much time I had to cook before the kids get grumpy.
We did some backyard remodeling. With unlimited funds and no HOA restrictions, maybe I’d get that Swiss Family Robinson tree house.
Heck, if it’s unlimited funds we have, am I still even writing this? (Probably yes.) I move to a giant plot of land with a tree house that amazes even the Swiss Family Robinson and start distributing money MacKenzie Scott-style.
I guess it was not written in the stars.
But, some constraints are good, such as considering data format to narrow down your chart choices. Just don’t let it be the only consideration.
Talk next week,