For information on controlling the heating level based on environment sensor data you could have a look at these publications for theory:
Ionesi et al
Rhee et al
Please note I am not sufficiently trained in heating theory to be able to judge their quality, but usually European work is fairly good
The amount of heat you will have to put in to efficiently keep your room warm will depend on outside temperature and wind and sun levels.
A heating curve (outside versus inside temperature), is building specific. To get one I would suggest simply collect data of average weather conditions while keeping your room at one temperature: Heater on-time, and outside temperature, wind force, and solar force. Best would be to have the extremes of the range of values you could expect, and about 5 datapoints in between. Then average the temperature datapoints and fit a curve (could be a straight line if you’re lucky) through the outside versus inside temperature datapoints (in for example a spreadsheet app). Don’t use datapoints with extreme wind or solar force.
With the fitted line you can then find the required on-time of your heater for every outside temperature, for that specific room temperature.
You could adjust the heater on time up by x % with more wind or higher room temperature or less sun, or vice versa.
Note: With hot water circulation the most efficient is to adjust circulated water temperature according to heat demand instead of on/off modulation.
But if you’re using direct heating from a burner with efficient heating (exhaust does not waste a lot of heat) this does not apply because it would just deliver X amount of heat per kg of fuel and you can probably ignore losses, so on/off modulation based on required room temperature would already be efficient.
So Question: do you think the current heating curve is the most inefficient element in your heating system?
Hope this helps