Restoration has been my first pick in every game so far (+1 production). After that, it really depends on the game. On Hard, you suffer from so much starting unrest that Civics is needed as soon as possible.
If there are Lvl 5 heroes in the vicinity, I think getting Heroes from the magic tree as soon as possible is vital - the AI doesn’t chase goodie huts/quests, but they will hire any Lvl 5 heroes they find. The same applies to Lvl 7, though that’s more long term. I’ve never seen the AI recruit Lvl 9 heroes. Apart from that, I pretty much don’t bother with the Magic tree. Poor cost/benefit, IMO.
On a game on Challenging that I didn’t finish, I actually went pretty hard on the Civilization tree, and succeeded in building all the wonders before the AI. On Hard, I was distanced so easily and badly on tech that I never had a chance (most of the wonders got built before I even got the tech to build them), so I went fairly easy on Civics and worked on Warfare a lot more. IMO, the critical Civilization techs are Restoration, Civics, Mining (+1 Prod), Trading (Roads… so critical for fast response) and Cooperation (4 man units - a 33% power increase + you can then research economics for that army size increase).
On warfare, I just beeline for Longbows as fast as is at all possible. And if I can avoid it, I don’t build any non-missile units other than scouts before I can complete the Armor/Weaponsmithing/Longbow line. Though in my case, I need war colleges too (for Athican longsword - which is just such a superior weapon to what comes before it that it’s ridiculous). Boar-spear armed light cavalry with Charge can be used to good effect, though. Most of the early units one can build suck terribly compared to what is available just a little later, which is why getting that early unit of archers from a refugee camp is worth its weight in gold.
Building wise, I simply calculate whatever building is going to generate the most production (i.e., by adding production bonuses or removing unrest) and go for that. Production is all important. My starting city build a lumber mill, and then pretty much spams Pioneers after that for at least the first 30 turns. Since spamming cities is such an effective tactic, growth is extremely slow, and consequently food is not much of an issue. And since I’m running with no taxes for much of the early game (except to keep solvent), financial building are completely pointless (doesn’t matter whether you’re getting 0% of 1 gildar or 5 gildar).