What am I supposed to analyze in "analysis?" What's the point of game analysis anyway?

Look at the game you posted.

Move 3: the computer blunders the knight, but you don't take it.

Move 4: the computer attacks your bishop with a pawn. You could take the pawn. You could move the bishop. You could take the knight that is still hanging. Instead, you make a totally random move that does nothing.

Move 6: the computer attacks your bishop again with another pawn. You could take the pawn. You could take the other pawn. You could move the bishop. Instead, you again make a totally random move that does nothing.

In that game, almost every move you made was a blunder that even beginners can see as a blunder. You seriously should check those "new to chess" lessons, starting with how the pieces move.

I'm new to chess and have posted about some of my woes in this forum. 

Playing the bots is ok but they don't make "human" style mistakes.  In the game you posted the bot leaves pieces hanging that you could've captured with little to no damage to your own game. I don't see many 700ish rated players making too many mistakes like that.

I would advise taking your time with the bots. Like someone else said, because they move quickly you feel under pressure. Remember against the bots there is no time control. Sometimes it can help to play a long game against a bot. Go away, have a break and when you return to the game look at it with fresh eyes as though it's the first time you've seen the board.

I've been playing 30 minute rapid games against people. This is a long enough time control that you can think moves through. Although sometimes I can make a real blunder of a move, it's those blunders that generally cost you and not your general development or what the opponent is doing. 

The puzzles and lessons on this site are also very handy to learn from.

Оh, such good question!

I could not understand that also.

Master Ginger said in his chess video lesson that we should not trust computer analysis too much. Human logic and machine logic differ.

During a game, I always try to check these things for a move:

- Is this piece defending anything? Will the defended piece still be defended after I move this piece?

- Is this piece threatening anything at the new spot? (capture or check)

- What pieces are defending this piece at the old and at the new spot?

I check these things for my moves and for the opponent's moves as well.

Analysis is a nice tool, but honestly, at beginner level, it doesn't give me a lot. Sometimes I get checkmated out of the blue, and I have to go through the steps to see where the mate was coming from. That helps. The tool doesn't always show me this. Sometimes I actually have an idea, and it works, but the tool identifies it as a blunder. Sometimes the tool identifies something as a blunder, and it shows me the best move, and I simply can't see why it would be so good.

The fun thing on a beginner level is that we're all beginners, we all make blunders. The Analysis tool is not considering this, as it is only a machine. It shows you the ideal solution. Real (beginner) people won't play that way.

I think you have that pretty much correct.  One should always self-analyze first.  We don't need an engine to tell us we hung a piece, or got a surprise mate/fork from our opponent.  And, beginners don't need to know that while this move was +0.20, some other move was +0.32.

Nevertheless, there are some things that I completely miss that the engine will find for me -- namely: great moves that I *could* have made (or my opponent could have made) and didn't (a fork, pin, skewer, etc.).  (I also use the engine to look at my openings: what was the first move that I went "off-book"? and "what did the masters in the game base play at this point"?)

For a beginner it helps in odd mating patterns and positions where you are stuck such as breaking out of a closed position.