HA data analysis and heating system


I have been using HA for a couple of years now after using domoticz for a few years. One of the main features used is the management of my heating with HA and the Schedy app from AppDaemon.
Apart from a few problems related to Z-wave, I am quite satisfied. However, like everyone here, I want to go further.
In Schedy, hours are set at which the heating is turned on and off. However, turning on the heating at 4 p.m. every Monday because you come home at 6 p.m. is not always cleaver.
Sometimes it takes more than two hours to reach the desired temperature and sometimes less. In the first case, when getting back, the house is cold and in the other, we will have heated more than expected, which is a pity financially and ecologically speaking.
For this reason, I wanted to use learning techniques on the data I recorded to try to calculate the time required based on the current outdoor temperature, the current house temperature and the target temperature.
Unfortunately, I messed up a bit with my data and I have very little of this winter, which gives results that are not always satisfactory.
So I wanted to know if:

  • You know other projects similar to mine. I searched but did not find. Maybe I looked wrong.
  • You have richer data than mine that I can use to train my models. Of course, your data can be anonymized.

Do not hesitate to ask questions, I will try to answer them.
Thank you very much to the whole community.