I'm not sure why you'd want to include AirBnBs and second houses in your measurement vacancy measurement — they aren't available for rent.
Sometimes people refer to underutilised properties rather than vacant properties for this reason.
I am very interested to know what happens to the proportion of underutilised properties during an accommodation crisis. The most recent linked chart in this thread is very interesting - what caused the steep increase leading up to the GFC (i.e. during a boom - often associated with a 'shortage')?