I understand the idea behind the principle, but where we have limited space, I personally prefer using that space to show the full difference between the lowest and highest values for the period. It allows the user to easily see at a glance which days have lower values, even if the absolute difference is very small. The alternative, where all values are absolutely close together, is to waste space showing the distance from 0, making all bars appear around the same height.

For example, if my weight were around 80kg, and decreased by about 0.6kg throughout the week, with the current graphs having a maximum height of 50 pixels tall you’d not be able to see any difference between 80.0 and 80.6 — all bars, including any in-between values, would be 50px tall. If I lost 1kg in a month, I’d be able to see a 1px difference in the height of the bars, which is not very obvious. Personally I’d prefer to be able to see my relative downward progress here, rather than being able to see that the absolute difference is not very large at all.

That’s using just one attribute as an example, but I think it makes sense for most others too.

On the other hand I don’t have any objection to changing the area graphs for averages to simple line graphs.

Yes, that’s why I suggested line graphs! That way the y-axis can go from the lowest to the highest value, using all available space, but without making it look like (to use your example) your weight decreased to 0, or that you lost half your weight when you went from 80.6 to 80.3, etc.

Ah sure. The problem is it’s not technically feasible for us to use line graphs in the main places we use bar graphs, like the web dashboard or the mobile progress tab.

Not super related, but this got me to dive into how Google Fit graphs data. It zooms out an extra grid-line on the scale to give you some context. So if your dataset has a minimum of 90 and max of 170 the graph axis shown will be between 80 and 180.