In retrospect:

Balancing the game was the hardest part, and the pacing isn't very good.

That's probably because I used too many unconventional mechanics, some worked better than others.

I planned things at the beginning, but it wasn't enough and I should have refactored my code so it's not 3000+2000 lines.

I should have used classes and the observer pattern.

I should have put most of the data in a spreadsheet and exported it rather than editing JSON directly.

It was a good idea to take heavy inspiration from (copy) the Kittens Game UI since I wouldn't be able to make anything better.

Saving by serialising one entire state object (and storing anything that shouldn't be saved separately) was a good idea.

The scope of the project was reasonable, maybe a bit too large.

The objective of telling a story was achieved, but it doesn't work as well as an idle game.

Drawing was fun, but I have no talent for it and it's time consuming, so probably not worthwhile.

Using a kanban board helped immensely with development, but it's not so useful when planning ideas and I'll stick to regular notes for that.

It felt worthwhile to finish the project, but I started to hate what I made after showing it to others even though I only got constructive feedback. Therefore it was wise not to show work in progress.

There's a bug I can't reproduce so I can't fix it. Some people found that they got too many extra resources when loading a recent save.