I first heard about sparklines when I was trying to find a way to display multivariate data while doing undergraduate research with Doctor Kody Powell. We were working on a novel method for finding faults in large chemical systems. Our work ended up getting published just recently in a journal. I’ll post the paper in a later post.
We made a simulation of a chemical plant with 400+ “sensors” that we introduced simulated faults to. We monitored these variables and tried to “see” the fault and understand what caused it, only through data means. On the post that I post the paper, I’ll go into more details.
Regardless, the premise of the paper is control charts are good for a few variables but if you have a lot of variables, displaying lots of control charts is very hard. Visualizing and interpreting that much data is so hard to do, especially with typical square plots.
I was researching ways to display lots of data in small spaces efficiently. During this search, I found sparklines and quickly fell in love with how much information these little lines could tell you in such a small space.
I’ve been reading Edward Tufte’s books recently and really fallen in love with his work. He’s made some amazing visualizations. I especially liked his section on spark lines. He shows this example about baseball to illustrate.
He shows so much information in such a simple way. We can see the cumulative wins of each team so we can see when teams were hot and when teams were struggling. We can see how many games over .500 they finished. We see their final record, home and away, and percentage. He also split it up by which league they are in. This is vital because if they did the whole league in one image, it would probably be too much.
At the top of the picture, you’ll see a mini sparkplot With just ticks up and down. This shows how a team‘s record goes with time. He’s able to smash a lot of information in a small space.
I love Sparklines.