Sleep as Android is now all about your Sleep Score and about how to focus on the 6+ different dimensions of your sleep and keep them in balance.
With the latest redesign of the Stats, Graphs and Sleep Score we wanted to even further focus on the Sleep Score and its connection to each individual sleep record you have.
This is why we created a unified visual language to be used in the Sleep Score as well as in the Graph detail screen. The common element is the pie chart – but also each dimension got an consistent icon and consistent way of showing the figures across the app.
How do you like the new look of Sleep as Android? Please let us know what you think.
10 thoughts on “Sleep as Android goes all Pie”
Fantastic to see deep sleep duration and a quick way of identifying details. I’d very much like the other states in a similar representation, esp REM and to a lesser extent Awake. Is efficiency duration inverse awake?
Hello Michael, many thanks. We are still very much cautious about any REM-based statistics, as REM is ~REM – just an estimate.. we expect REM to be this in healthy individuals but we do not directly measure REM.. regarding awake.. we have updated to Efficiency pie based on your suggestions.. so now it shows Efficiency % – the radio of sleep tracked vs. actual sleep duration (without awake).. but the duration underneath is in fact awake duration, so I hope this gives a quick way to check on awake..
I like the reprioritisation! (Including efficiency – is that a new metric?) Really exciting to see you update this. It’s an interesting, compact design.
My first impressions and stray thoughts:
– More obvious icons may be needed. E.g. the “irregularity” and “duration” icons are not clear to me. Ditto “deep sleep”. Duration could be an hourglass, stopwatch, or a clock icon. Deep sleep could be a pictographic representation of a person cupping their hands under their head, a pillow, or nightcap (hat), etc. Irregularity could be a clock set to 3am, an exclamation point / warning symbol, a more pronounced/jagged line chart, a person yawning, etc.
– Something strikes me as off about the labels/titles. (Maybe the spacing?). Perhaps they could be hidden until tapping one reveals all of them in-place. (Related note: the font colour for labels inside the sleep score box is quite low contrast. I can understand why – and it’s legible to me – but some users might struggle).
– It isn’t immediately obvious to me whether the various scores (apart from duration) are good or bad.
– Similarly, the meaning of the pie chart value isn’t clear to me. But perhaps that’s because all of the metrics in the screenshots happen to be very good? The pie chart lines also take up quite a lot of space (though I also feel the lines could actually be thicker!) and the spacing between their lower edge and the labels is ever-so-slightly inconsistent. A simple left-aligned vertical bar (ideally with colour coding on the positive or negative segments of the bar) might be a better use of space.
A slightly broader range of positive/negative colours might be helpful here. E.g. either a 4-step red/orange/yellow/green indicator (that might clash with your current colour scheme), or colouring the negative/deficit part of the pie chart alone. (E.g. -12.5% duration = grey, -25% = orange, -50% = red).
Is the level of deep sleep good? Bad? This is perhaps the most important metric in the app, but it’s always been very difficult to understand.
Two numbers are given (duration and % share of total sleep), but neither number has a comparator or benchmark. (The user would have to know in their head what a good number should be). Speaking as a long time user whose overall sleep duration varies a lot, I find it very hard to parse the deep sleep numbers in the current UI, do the small amount of math or recall necessary to figure it out, etc. (2 hours of deep sleep might be a lot one day, and very little the next – relative to overall duration).
Ideally, this (and most) metrics should be benchmarked against the average over a defined recent period. For example, against the 28-day average of the previous month (or 7-day average of the previous week) of deep sleep. So instead of:
35%
2:34
You could show:
35 <— Deep sleep %. % symbol removed to indicate this is a score (which also happens to be a %).
+5% <— 5% absolute increase on benchmark (e.g. current value less the average of last 28 day's deep sleep %). (Side note: of course, maybe there's a point at which too much deep sleep is a bad thing!).
Ditto with snoring. If deep sleep was changed as above, then ideally the Snoring numbers would be styled as:
3 <– Snoring % or score. % symbol possibly removed to remain consistent with other on-screen metrics/scores (though maybe keeping it would be more intuitive, since lower = better unlike the other % metrics where higher = better).
-10% <– 10% decrease on current benchmark. (Negative values green, and again % value is absolute rather than relative).
Hello @ric_cde, many thanks for the details. Please see here https://docs.sleep.urbandroid.org/sleep/sleepscore.html .. each dimension in the score has an healthy and unfavourable threshold.. so the progress is any progress between those two thresholds.. if you reach healthy range progress is 100%, under unfavourable it is 0%..
I will review your feedback one by one and see what we can do. Also we released several updated to the new look in the mean time (sorry for a late response, but it was a very long feedback :)).. so some of the issues may be already addressed?
What a pity that you have dropped the text advice “For optimal deep sleep/ rating go to sleep at / sleep at least”. That was crucial for me to determine the best time to sleep. Please bring it back!
Hello Andreas, sorry about that we have dropped it for a white to manage all the changes, but now it is back and I think in a much better shape.. We have added efficiency to the mix – I think efficiency may be even more interesting than the other measures in this context, but more importantly we have fixed an over-fitting issue which could have matches ridiculous curves on the samples.. So now we are looking for smallest RMS for the regression by favoring simpler models.. So in many cases you will just see a linear model – but this is a good think as this respects the data better and only if a polynomial model really makes sense we use show it… Hope that helps..Petr
Hi there – where can we find the SDANN?
Hello, you can find SDANN in Charts > HRV – this is the development of SDANN for the whole night, where HRV Before wake tracks development of HRV which we measure in the last calm period before awake.. Hope that helps..
Thank for your reply Petr. Before the update the SDANN appears in the morning summary but I can’t seem to find it anymore. Am I missing something?
Hello Cathy, after the update with the pie charts we call it HRV Gain and shows HRV after sleep and the gain (HRV after sleep – HRV before sleep)..HRV Gain is shown as the other pie charts on the Sleep score and on the detail screen of graphs – maybe you just need to scroll if you have more pies?
Than we also calculate overall SDANN an an average from all the values during the night and this is called simply HRV and is only available on charts as too much HRV measures started to be confusing..
Hope that helps, if case you still can’t see it, please contact support@urbandroid.org with a screenshot..
Many thanks..
Petr