I wanted to write a bit about supercities.
Supercities are very important for small civs, and deity pangeas often leave us players with such small civs.
The way to keep up in science then becomes the supercity.
A supercity's principle task is to generate lots of beakers. However, bulbing great scientists is the most effective way to research in early to mid game. That's why supercities often end up as being both GP farm and science city.
This is my capital in the Spanish game. Palace + good starting positions usually make capitals the best supercities.
The foremost thing Madrid does is produce great people.
However, it also runs a few cottages (actually more like villages already). They are build in the best possible locations, namely the river grasslands. The reason to keep them running instead of, say, four more irrigations and two more scientists lies in the mechanics of bulbing. As powerful as bulbing gpp-beaker ratio is, most techs cannot be bulbed optimally. Many are too cheap to fully utilize even one great scientists, and all lose a fraction of the bulb when finished with it.
That's why it's good to have traditional research power.
In a larger empire, this can be achieved by building research when needed. In a miniempire, a high commerce supercity under bureacracy and academy is best.
It's very important to grow your supercity. Monarchy is great for the happy cap. And for health, don't underestimate forests. Chop them only when you actually have the pop to continuously use the tile beneath, or when you absolutely need to rush crucial wonders.
Don't build workers and settlers in the supercity once you have a few other cities going. Focus on growing it as large as possible. Don't build units you can build somewhere else. Focus on advanced infrastructure like aquaducts, harbors, etc.
As it grows, some buildings become more important then usual. Monasteries are one example. Research 10% more doesn't sound like much, but in a supercity it may well be a dozen bpt.
Normally the supercity works only resources, cottages, irrigations, specialists. Improvements like windmills or normal cost tiles are totally suboptimal.
Mines however follow different rules. They are used some of the time, when the supercity needs to build something fast.
Also a thing of note, I sometimes overlap to-be-cottaged tiles in other cities bfcs. This way, the supercity needs no bother working cottages/hamlets and can focus better on being a GP farm. Once the cottaged tiles reach 4 commerce or so (that's effectively 6 with Bureaucracy) the supercity starts working them.
National Epic is the only must-have national wonder. Oxford seems like a good second, but that depends on various circumstances.
Finally, to balance the supercity's peacemongering focus, the rest of the empire can develop as more of a production base, avoiding cottages for after successful expansion wars and emancipation.