I saw a kinda clever schematic for a crude air flow sensor using an incandescent lamp and a constant current source, with some associated analog electronics to measure associated changes in the filament temperature wrt airflow velocity.
Problem is that the lamp needs to be cracked open and the thin filament exposed directly to the air for this trick to work. pretty fragile and impractical for high velocity air. Maybe use a very small (intact) lamp like grain-o-wheat and look at the envelope temperature in the infrared. like with a wireless mouse IR camera or IR photodiode