I’ve been thinking a lot lately about visualizing error. Standard methods like error bars, or shading a range don’t quite seem to capture it… I find that many people mentally draw a line right down the middle and putting less importance on the error.

So here is a pass at a different way of visualizing time series data. This dataset shows 3 trials with identical conditions run sequentially. I find that the motion of the points speaks to the variation and the shading in the back provides a good context.

You can see the code and a live demo here