Graphs in the Oura app

Poor-quality graphing of trends is a problem in every health tracking app I have used with my iPhone, including Health, Activity, HealthView, AutoSleep, Fitibit, HeartWatch, and a few others that I deleted after a few days' use. Oura graphs are good for many variables, but not for all.

Below is an example of a good graph from Oura (Graph 1). The bar used has a meaningful color gradiant. Both Axes are clearly labeled. And, most important, the scale for the vertical axis is based on norms, so that small and large differences can be seen at a glance and the information conveyed can be used meaningfully.

Graph 1

When a vertical axis is auto-scaled based on the data, small differences are exaggerated and impossible to understand visually. This is a very common problem in Health apps. Graph 2 (below) is an example of a bad graph from Oura. Nightime respiratory rate typically varies in adults from 12 to 20 respirations per minute. Mine has been close to 15 every night for 3 weeks. But in this trend graph, the vertical axis ranges from 14.5 to 15.5 making very small changes look big. At a glance, the graph is alarming. It looks as though my respiratory rate has been dropping dramatically over the last 3 weeks. If this graph had used a vertical axis of 12 to 20 (the normative range), the changes over the last three weeks would have been barely noticeable, communicating at a glance that this is a stable variable for me. [By the way, the 6, 7, 8 labels on the horizontal axis refer to Weeks 6, 7, and 8 of the year, corresponding to the weeks from the February 3 to February 23, 2019. It took me a few seconds to figure that out. Kudos to Oura for labeling the axis clearly.]

Graph 2

The Oura Readiness graph is particularly troubling. I would guess that users of the Oura ring look at the Readiness graph every day and try to use the information to plan the day. Here is a recent one from my data (Graph 3). 

Graph 3

You can see in Graph 3 that my resting heart rate has varied from 46 - 53 over the last 5 days. If the graph had used a norm-based range on the vertical axis, from 0 to 100, these differences would appear much smaller and the information conveyed would be more meaningful. At a glance it would be clear that my resting heart rate has not varied significantly.

What makes Graph 3 (above) useless is that there is no label for the scale on the vertical axis, making it appear that my readiness scores have been very low for 5 days, also in the 40s and 50s. Actually, they have ranged from 65 to 82. If both the resting heart rate and the readiness score were plotted on a scale from 0-100 and the vertical axis were labeled, a different pattern would emerge. The blue bars would be higher than the resting heart rate everyday and it would be clear that while my resting heart rate is in the healthy range and does not vary much, my readiness score varies substantially and has not been close to what it should be.

Below is an Oura graph that I can’t figure out how to use (Graph 4).

Graph 4

Graph 4 is supposed to tell me the intensity of my activity over the previous 5 days and today (which was just starting). It is true that yesterday I had a full day of only low intensity activity. I was working at my computer all day. However, the length of my day was the same as it was on the previous two days. So, why are those bars so much taller? The day before yesterday I went for a brisk walk and so had some medium intensity activity in a day that was also spent at the computer, but the day was no longer than yesterday. The graph would be more useful if it were divided into inactive time, low, medium, and high activity time. The height of the bar would not vary much, because the length of a day, does not vary much. The bar would be divided into each level of activity, so on days when there was mostly inactive time, that segment of the bar would consume most of the height of the bar. This is exactly the way in which the stages of sleep are graphed (Graph 5). Activity should be graphed in the same proportional way that sleep is graphed. Graph 5 could benefit from a vertical axis, but even without one you can see the sleep pattern at a glance.

Graph 5

The Oura app is regularly updated, so these issues may be corrected in future versions.