Are Sparklines Smart Lines?
by Stacey Barr |They’re so hot right now, sparklines. They thankfully bring some sanity to performance dashboards (goodbye goofy gauges). But they do have a limitation no-one seems to be talking about.
Sparklines are the crafty invention of statistician and data display guru Edward Tufte, and presented to the world in his book Beautiful Evidence.
A sparkline is a tiny line chart stripped down to its essential function of showing the time series of a specific measure. They are so tiny they can easily be embedded into text, tables and other tight spaces.
Little wonder Stephen Few has embraced them in his landmark book Information Dashboard Design.
I like them because you cannot manage performance when you’re looking at pie charts and gauges and dials. You need to see change over time.
What’s more, you need to separate the signals of change from the noisy variation that exists in any measure’s time series. And that’s where I think sparklines meet their limitation.
Beyond sparklines…
Sparklines still encourage you to focus too much on the first and last values of your performance measure, rather than the true signals of change that are based on the patterns in the time series, not the points.
Separating signals from noise, and drawing your eye to patterns rather than points, is something XmR charts do well. XmR charts are a specific type of statistical process control chart, discussed in detail (very interesting and readable detail) in Donald Wheeler’s book Understanding Variation.
They are graphs with 3 very important features that make performance measures so much easier to interpret validly.
xmr chart
I wondered what would happen if sparklines courted XmR charts? So I played with it and ended up with something that I reckon absolutely rocks. You get a tiny graphic that works well in text, tables and dashboards AND it is easier and faster and more valid to interpret what performance is really doing.
I call them ‘smartlines’. (I would like to call them ‘XmRtlines’, but pronounced ‘smartlines’ – too cryptic though, I think.)
What do you reckon, Stephen Few?
TAKE ACTION: Take the time to learn about XmR charts, and when you build your dashboards, give the smartlines (or XmRtlines) a go!
Speak Your Mind
Join Measure Up
Sign up for the Measure Up newsletter and get free access to the "10 Secrets to KPI Success" online course and e-book.
Upcoming KPI Training
>> Africa, In-Person, Dubai UAE, 3-5 May 2023
>> Australia/NZ/Asia Pacific, In-Person, Wellington NZ, 9-11 May 2023
>> UK & Europe, Online Interactive, 22-26 May 2023
>> North America, Online Interactive, 29 May - 2 June 2023
Connect with Stacey
Haven’t found what you’re looking for? Want more information? Fill out the form below and I’ll get in touch with you as soon as possible.
167 Eagle Street,
Brisbane Qld 4000,
Australia
ACN: 129953635
Director: Stacey Barr
Great post Stacey.
We’re looking right now into integrating sparklines into our product. And in sitting in on some of Stephen Few’s trainings, like you said, he’s also a big fan.
But my favorite line of this post is
I like them because you cannot manage performance when you’re looking at pie charts and gauges and dials. You need to see change over time.
I would add columnar format reports to that as well. I am constantly preaching this to our customers. But it’s amazing how stuck people get in seeing the data even when it doesn’t tell them nearly as much as a well defined chart.
There is certainly a place for all of those (actually, maybe not pie charts) but I try to convince people not to get stuck on any one type of report. Find the right one that quickly conveys the data you need to see and drill down from there if needed.
I’m definitely going to check out XMR charts. We had something similar in an earlier version of our product and the mass population didn’t care for it, but maybe it was our implementation. I’m going to take another look at them.
Finally! approximaetly 80 years after Shewhart used statistical methods to assess meaningful variation in census data, and over 50 years since Dr W Edwards Deming made them familiar to manufacturing operations the world over, they stand a chance of getting into common usage. The other key message these charts provide is that you can only shrink the span between the control lines – ie reduce the uncontrollable variation – by changing the process which creates the data, not by adjusting the input to aim for a different output value. Perhaps ShewCharts might be an honourable mention 🙂
Russ, I love it: ShewCharts! Thanks for reinforcing this point. It’s so important and I find it very hard to get the message across to people. Not anyone’s fault but rather the product of a long history of looking at data as precise rather than variable. I didn’t word that very well, but I’m sure you know what I mean. 😉
Stacey, I like the term XmRtlines, but I agree it is probably a bit cryptic. More importantly the value of combining sparklines and XmR charts is spot on in providing the information you need simply and clearly – where you have got to, what variation there is and where you want to get to. Fantastic.
I am happy I have learnt about sparklines. Will that be helpful to produce a perfomance report “objectives versus achievements”? I like the XmR chart also.
I am looking forward to your improved “smart lines”. What method is beneficial to capture and interpret the performance.
Stacy, I have subscribed for PumP blueprint self paced on line course and looking forward to receive the material soon and bring a significant change to the subject and evidence good results. Looking forward to your support and assistance.
Interesting to read about your trip to Oslo, I wish you all the best.
Very best regards.
This is really a great feature. Haven’t used sparklines so far but I like your version of the XmR chart in sparkline.
Did you do this in Excel? Don’t know how long you played with it but I still haven’t figured out how you managed to get the measures, mean, upper and lower variation in one sparkline 🙂
Jim, I am sooooo glad to hear that you’re interested in XmR charts! I hope you do find a way to integrate them into your software – you’ll be one of the first (and I say “one of” only because I don’t know every software app out there for dashboards, but I have never seen them in any so far).
Thanks Peter – XmRtlines is fun for those in the know, but the idea is to get the concepts of smartlines out there, particularly the XmR chart concepts of meanline and limits of natural variation, to people who are not yet in the know!
There is more work to do on making them easy to set up in Excel and other dashboard apps, but I’m sure it’s possible with a little help from the experts.
Mala – if you’re interested in objectives versus achievements, I think this is the same as target versus actual, and that’s what the XmR charts (and smartlines in particular) can do with ease. Sparklines don’t do it because they encourage you to compare the most recent point of your measure’s time series with the target and that’s not statistically valid.
Do our performance measure comparisons need to be statistically valid? Of course! Statistics is the science of drawing conclusions from data. Isn’t that what performance measurement is too?
Jan, yes I did use Excel. I’ve uploaded my Excel worksheet here so you can see how I did it:
https://www.staceybarr.com/downloads/StaceyBarrSmartlines.xls
It was a manual process of setting up the XmR chart first, and then removing the axes and labels and then simply resizing it to fit into a cell in the worksheet.
Much more work is needed to make these smartlines work easily, and that will come with time. If anyone reading this is good at writing add-ins for Excel, please contact me at info@staceybarr.com !!
Thanks for your example worksheet Stacey.
It took some time before I realized that it had to be a resized ‘normal’ Graph.
Hi Stacey, I like the idea of Smartlines, did you pull these up in Tableau?