I always like to approach a new coding project by concentrating on the data model first. I've seen a lot of projects built on extremely convoluted database structures and it really hurts because it makes it hard to add new features to the project.

So I look at the requirements of the new project and try to come up with a basic data model. Then I like to think about what logical future additions to the project could be. And using those, I try to see if the data model is flexible enough to be able to handle those additions fairly easily or if complex migrations or hacks would be needed to account for new use cases and features.

I think once you have a solid data structure and database technology, planning out an API or rest of the software is pretty straight forward. I like to create reusable pieces of middleware early on in the project which makes it easy to apply consistent functionality with ease to different API endpoints.