i wrote some scripts that runs airodump ng, then clean the csv, send all beacons with timestamp to a database.
this helped me to collect mac adresses for some time now and allows me to have alerts if neighbours are back from vacation (e.g. using home assistant i get notifications ob androidTV)
i did not go for bluetooth because espresence for homeassistant does that out of the box.
https://espresense.com/
exactly that.
i am currently in the trouble of visualising the data (i use grafana) but still not happy.
what you might wanna do is just collect data and query it later to learn about the mac adresses you are looking for.
e.g. run airodump every 10 minutes, write mac and timestamp to mysql. then you only need to write down when your colleagues/neighbour arrived or left for like 2-3 days and then query like "which devices were present after time x on these days ..." and the larger your dataset the more accurate your results.
the guy in the peremeter detection has like 1 client a day ...so thats easy. but with more ppl around you, cars being APs etc. you need more data to be sure which mac is who.
maybe if you want to track cop cars use that and gpt4o to watch a cam feed of your street to match copcar & mac.
if you got a few solid hits you can use that to figure out patters...like you know cop1 is around everyX hours which tracked mac matches that.
i am sure mac is much more reliable than bluetooth due to range, consistency etc...