Edward Tufte would be happy ... bg Sparklines

Some of you may be familiar with Edward Tufte’s books or seminars on presentation of high-information-value charts and graphs. I’ve been a “fan” of his for nearly 30 years now and he even came and visited me at work once to see some of the tools I’ve developed. One thing he’s been pushing for a few decades now, is “Sparklines”, tiny almost phrase-like ways of looking at data. More about Sparklines here: https://www.edwardtufte.com/bboard/q-and-a-fetch-msg?msg_id=0001OR

I have been dissatisfied my whole T1 lifetime (nearly 40 years) at first with urine test logs and then bg log sheets. I am so happy that a few months ago I found the “AGP” tab under my Dexcom G6 clarity and it shows the data so much more richly and completely than any log sheet ever did.

While I occasionally come here and gripe about Dexcom adhesive or Dexcom presoaking, the guys who did clarity and the graphical reports did it right. Look below… that is 4 weels of bg’s, measured every 5 minutes every day. Or over 8000 pieces of data. All shown using wonderful Sparklines on a calendar overlay. Edward Tufte would be very happy.


I’m a big fan of any data graphical representation that enables treatment decisions in a relatively simple and quick way. I think the Ambulatory Glucose Profile Report fulfills that goal.

Now, I realize that it takes some thought and study to understand what the AGP suggests, but I think it is well worth it. In addition to the concise listing of four weeks of daily glucose traces, the AGP trace itself condenses all that data into a useful tool. While I prefer the 14-day report, the 30-day report can reveal some lessons.

As an example, here’s my 30-day AGP trace and what I think it’s telling me.

1.) I need to make some insulin adjustments to my evening meal program to nudge down some post-meal and early sleep-time BGs. I could adjust any or all of the following: my mealtime insulin to carb ratio, basal rates, and insulin sensitivity factors during this period. I use a pump, so these adjustments are available to me.

2.) The mid-day brush with hypo territory suggests I need less insulin and/or less aggressive timing for my first meal of the day. I’ve recently adopted a daily sauna in the late morning before I eat so that activity seems to increase my insulin sensitivity. I’ve found that simply decreasing my pre-bolus time from 30 to 10 minutes has helped a lot.

Can you share more of how Edward Tufte’s ideas have influenced your glucose management? Nice consistency in the displayed lines, @Tim12! Persistence is your friend.

1 Like

What did he say? :innocent:

Start with the data. And, I agree, “PowerPoint sucks”.

1 Like

Edward Tufte – “We want to be approximately right rather than exactly wrong.”

He seems like an interesting character.

1 Like