My latest slow game - Analysis and advice on how to improve needed

Below I highlight the moves that took me the most time (had trouble finding a good move):
Move 9 Bd3 took me 15 minutes
Move 11 0-0 took me 23 minutes (!)
Move 12 Qce took me 10 minutes

Unfortunately, after move 12 I only had 15 minutes on my clock... (This is common in my games, I run out of time very early). Hence, for move 20 I used less than a minute.

Any ideas on the game and tips on how to improve my game would be highly appreciated.

You should do your own analysis first, that way others have an idea of your thought process.

What is 6. h3 all about? My candidate moves were: Bd3, Bb5, a3 and c3 Not in any particular order, the bishop move means 0-0 is available, the pawn moves make the bishop go away. You are not getting mated on f2 due to Bg3. If the knight does go to g4, then kick it with h3 after you move king bishop. 

9. Bd3, what was wrong with Be2 especially with the building e5 dispute. Also in Dutch like games black tends to play ... Qe8, Qg6 etc.

11. 0-0 now me I do not like my king on the same file as their king because all the wood in the way has a habit of vanishing.
Now I know what you did, you looked at d4xc5 Nc6xe5 ... OMG what if he plays Nxd3/f3 check err do I c5xBd6. Hold on where was I, oh yeah d4xc5 .... now how about I start with NxNc6 hmmm don't like his BxB PxB  opps no he can't do that because his queen is attacked by my knight. round and round.

A way of not doing so many variations is e5 is attacked 2 times, defended 3 times - Without the pawn it is still 2-2 and that works for you. So you know d4xc5 does not unhinge a defended piece.
So my 12. is c4xd5 after a quick thought about Ne5xNc6. Your Qc2 could get a ... c4 reply

Your move 14. Qe2, I like that. It tells me you have seen ... Nb4

15. NxNc6, on principle. It means your queen is attacking (e6) and his queen has the painful duty of defending a pawn. What is my queen doing, defending a pawn has a sinking feeling.
15. ... Bh5 now this is annoying. 16. Bg5 is sharper as ... Nxe5 ; d4xe5 forks and pins.
Getting out of the pin with 16. Qe3, defending the bishop, sure.

17. g2xBf3 ?? You now have doubled isolated pawns and your king is exposed to weak squares ( remember h3). Either piece would have better, knight being my choice. 17. Ne5xBf3

Now we get to 20. You were out-combined because you did not see Bh2+, it happens to all of us.
Famous advice: If you see a "good" move, sit on your hands and look for a better move.
20. Bb5 maybe or super cool 20. Kh1 because the e-pawn is falling anyway.

23. Maybe c4, because d5xc4?? Bxc4 and you get the exchange back. Two well placed Bishops and a rook can give two rooks and a knight a hard time. It was unfortunate that grabbing the half-open file, normally good, was not so in the game because black could force a rook exchange.

does the bishop typically come out to f4 like that? i would have played the natural c4 and controlled d5 with a pawn.

While others have done a good job of looking at your moves and offering improvements, I would like to look at your thought process.

In particular, your time handing in the opening suggests that your thoughts are jumbled.  You either don't know what to focus on, you are experiencing nervousness/doubts and triple-checking every variation, or you are unnecessarily calculating at length when you don't need to.  Perhaps it's all three, perhaps it's something else.

If I asked you what we should do in the opening, I'm sure you could tell me: develop all our pieces and castle.  On move 9, you spend 15min on a position where you haven't castled and have one piece left to develop.  The only thing you need to think about it developing and castling.  That's it.  On b5 your Bishop doesn't do anything, so decide between e2 or d3.  Do a quick blunder check, then make your move.  Anything more than three minutes here is excessive.

Move 11 is similar.  You know you need to castle, and you can do it immediately.  Your only question should be: Kingside or Queenside?  Your c-pawn has moved and Black's c-pawn has moved forward, so the Queenside prolly isn't the safest.  Let's go Kingside, especially since Black has already castled.  Do a quick blunder check, but everything is defended and nothing is critical yet.  Castle.  Again, three minutes max, and hopefully even less.  Even if you were thinking about what to do next, which is important, you know you need to castle, so castle.

Your twelth move IS a good time to think, because your development is complete and you need to decide how you will move forward.  It's important to think not just about one move, though, but to think how you want to play in the middlegame.  10-15min is a good amount of time, but this assumes you are thinking about planning, not calculating and re-calculating variations.

One of my biggest stumbling blocks was trying to make sure every move was safe.  I would calculate as far as I could every move ... but that took forever, I got tired, I was now down on time and my play got worse.  I learned to get rid of that, or at least get better at it, and it made a big difference.  If a position isn't complicated, such as an opening where we need to finish development, just do that and then start thinking.  You know you need to castle at some point, so do it now, that sort of thing.  The clock is an important element of the game, just as important as space or material, and I think learning how to deal with that will help you going forward.  Give yourself permission to play obvious moves and see what happens.

Move 17.gxf3 was the worst move. without a concentration of powers on 'g' file how can you get it opened and make your king vulnerable.A lot to learn about the Q side pawn moves which were below average in a position in which you had a better pawn structure on the queen side.

On move 31.f4 is positional blunder, you could still up a fight with 31.Kf1 follow up Ke2, so you can defend the queenside pawn, even in a bad position make it hard for opponent to win, a lot stiff resistance save a game from a lost.

Yes. "positional blunder" means have a good look at the range of the black squared bishop. It is sitting on h2-b8 (do not block it out) On f4 it pesters h6 that has to be defended, blocked(g5) or moved.
Your other bishop can go on d1-a4 diagonal, pushing on the d-pawn.

Imagine at 31. you could pick up Bd3 and drop it on d5, do you see how strong the pair would be.
Many authors have suggested look at imaginary positions and then see if they can become reality. 
Certainly playing above 1080 for sure.

Good inside and explanation, especially on "imaginary position".