One thing I noticed about the load cell: It is not fully temperature compensated. During a 24 hour period you can see the zero (tare) point move up or down by 1-2 grams. This is not a big issue since a typical bird for this type of birdhouse weighs about 10 grams but maybe there is something that can be done here.

I was thinking about using the temperature measurement to compensate for this drift. But the CPU thermometer used right now is not very well suited for this kind of correction. The CPU temperature is always a few degrees higher than the ambient temperature. And with the shunt regulator also close by it's not measuring the actual temperature of the weight scale.

For the next version of the birdhouse I was thinking about maybe mounting a Dallas Semiconductor DS18B20 one wire thermometer next to the load cell. That thermometer could help level the temperature drift and would also provide a better ambient temperature reading for the birdhouse.