Making a Sparkline Better

I’ve always admired smart data visualizations and especially the designs inspired by the world-renowned author Edward Tufte in his works such as Envisioning Information. If well-designed charts are the language of information understanding, Sparklines are the symbolic equivalent of Blink - those snap judgments we often make that can be just as good as—or even better than—the decisions that we make when we analyze a situation carefully.


In this sparkline experiment which was originally inspired by some internal tasks at Stream It and also an earlier post by Paul_Danyliuk, I wanted to create a sparkbar with one added twist - a sprinkle of machine learning to predict where the spark bars would be in the future.

As I envisioned a way to do this in my mind, the Pack was extremely simple, but that should be a red flag for anyone who codes for a living. :wink: In reality, there’s a lot to think about when plotting and coloring positive and negative values and then predicting what it might look like into the future. Simplifying how my team uses this is a big challenge - no one wants to write complex formulas to achieve data visualizations.

I took some random company names and data and threw it at this new pack and it seems quite useful. I think this one will be my first contribution to the Coda Gallery. And now it’s available.


Look forward to it @Bill_French