Its a bit of a toss up in my mind. Sense has been well funded and appears
to have some good algorithms in existence or growing along with a growing
user base. This along with a good deployment of machine learning should
make their platform keep getting better. That said I have not found a real
good way to train their system.
Neur.io has a much more open approach with APIs to access data and thus the
ability to integrate it with other platforms and tools. I have not found
any APIs for Sense so far.
In terms of the quality of the dissagregation, I have not spent enough time
to get either working well in my house. Of course this is just the point -
if it takes lots of time it is not a great solution. In their defense, my
house is pretty complex, with multiple fish tanks running lots of pumps and
heaters and 2 heating / cooling systems thanks to an old remodel.
Both were easy to install and I did then myself in less than an hour each.
they operate fine side by side and the data seems pretty consistent when I
look at them at the same time, though certainly not exact.