Hi @miracj, when I saw your post I initially thought "of course they are proportional!"
On second thought it’s probably not quite as simple. I guess we can agree that the bubble size is a way to depict relative power consumption - the bigger the bubble, the more power a device consumes (or produces - in case of the solar bubble). You probably also noticed that during “quiet” times, low-power devices/groups of devices are depicted larger than when lots of devices are drawing.
Earlier this morning my Sense app was only showing Unknown (~360W), Always On (300W), and a small fridge (107W). The three bubbles were sized to fit comfortably in the space available on the screen. Then my geothermal heat pump kicked in (5800W) - and its bubble initially didn’t even fit on the screen. It took a moment and then all bubbles were resized to fit on the screen again. Of course now the 3 initial bubbles were much smaller - but you’ve probably seen this already.
Now, what about proportionality? It looks to me as if the lower power devices are getting relatively more real estate than the high-powered devices. Otherwise, in my example above my fridge would have been a tiny dot next to my heat pump, and I probably wouldn’t have been able to read the label anymore. In other words - yes size is proportional to power, but it’s weighted towards the smaller devices so we can still see them.
With my interest piqued, I was trying to figure out an actual formula. I took a ruler and measured the diameter of the circles and dropped them into a spread sheet - alongside their respective wattages. Then I compared the relative ratios (each device’s wattage in % of total current wattage, and also in % of wattage of largest device) and compared it against the area of the circles as well as the volume (if they were actual 3 dimensional bubbles). The volume comparison actually came the closest, but was still a bit off. Maybe the Sense developers will let us in on the secret at some point. But I can probably live comfortably without ever knowing the exact formula…