The tactical bells.
When should you calculate more than usual? That is a question that even a GM cannot properly answer, but tactical bells are the number one pick.
The idea is simple. If you check for tactics, you will find some at times. If you check most tactical puzzles, they have a pin, weak backrank, overworked pieces, etc. Look for that in a game. If you see a weak backrank, calculate. If you see a piece that has two jobs, start calulating. A pin? Try pawnstorming!
In attacks, always calculate several moves ahead. That is the only way to survive the piece sacrifice you might have made.
The positional whispers.
Put a knight in the center. The bishop pair. The pawn structure. Look for those. If you think that you can put a knight in the center, can they push it back? Will you force him to give a bishop up for a knight?
The endgame drums.
Endgame is the part of a game where you should calculate the most. Less pieces, more calculation. If you are in a king and pawn endgame, check for everything.
On everything else. Advancing a piece, pushing a pawn, etc, don't need a lot of calculation. The more chess you play, the more you will feel moves like that come right to you. Those moves need less than 10 secs of calculation.
In the end, playing reveals more. The more you play, the more you do. Hope it helps.