oem/habitat-ml

fighting humidity with machine learning and iot

HABITAT

Fighting humidity with Machine Learning

motivation

We moved into a new flat a while ago and while the flat is great, we kinda felt that it was more humid then it maybe should be.

So this little project happened. It visualizes and predicts humidity values (the counterpart being a sensor that measures the humidity).

That way we can for example see what effects opening the windows or using a dehumidifier have. And if there is a problem at all.

Ok, sure, but why would I care?

Maybe you are planning to do a little project where you combine iot stuff with machine learning. If so, this might help you with some of the basics, like how would I serve the model and visualize it? How should I structure my code best?

Also included are some notebooks, exploring the data and trying out a few models.

You can check out the running webapp at https://habitat-ml.herokuapp.com

The initial loading might be slow, since I am using the free dyno on heroku and it might take some time to coldstart. The actual webapp performance it pretty descent- it is not the app, I am just cheap ;)