I don't think it's a great way to learn, tbh. The computer analysis will tell you if something's good or bad, in a black-or-white kind of way, but it doesn't tell you why it's good or bad. Also, computer analysis tends to be better from a tactical point-of-view than a strategic/positional one.
E.g. if the computer analysis suddenly goes from +0.7 to -2.4 then it's a good bet you just blundered a piece, but if it goes from +0.7 to +0.4, it's often not that easy to figure out why.
Doubled pawns aren't always bad, btw. There are many situations where they can be quite handy.
Let's assume I am way too lazy to go about reading books or pursuing a more structured approach.
I know the basic principles - control center, connect rooks, avoid doubled pawns etc. Now if I simply read computer analysis of all my games, will I be able to get a hang of more advanced patterns? Any major downside to this approach?
Also, what's the best analysis software to use?
Thanks.