I would suggest you start by just coding a text-based game. Yeah, you heard me right. Maybe a text adventure or whatever. A D&D game printing out damage as logs. FTL clone, with text descriptions only. Anything. You can create a surprisingly engaging game just with basic standard input and output - even multiplayer games (the MUDs were basically this). Print a basic ASCII map, and now you can do Nethack.
The reason for this is: we all want to make beautiful AAA games. But if you have no clue where to begin, it means that you need to develop your intuition for the game logic first – otherwise, you would probably know what to look for :)
If you start by downloading Unity or similar, now you'll be bogged down trying to learn all its systems (without a clear understanding of _what_ you need to learn and what you can ignore, for now). You'll also be bogged down by the need for assets. Sure, you will have a full blown 3D engine, but it's still incredibly boring when all you have is a bunch of cubes or premade assets, so you are right back to square one. Only with more complexity. A lot more - the more visually complex the game, the more code you will have to write that's only concerned about visuals. Getting a character to, say, swing an axe and make it look right and that it is actually hitting something involves a surprising amount of work. Yeah, you could use existing sample games nowadays, but that isn't really teaching you much.
Then it depends on how much background you have. If you were, say, a front-end developer with any experience, you could use that to add some basic visuals. Think Tetris. You can do a lot with very rudimentary tools, as long as everything is kept simple.
At some point, you might _need_ to display graphics (maybe that's the whole point of your game idea). I say "might", because Dwarf Fortress, which is the subject of this thread, never really did. In which case you have some more decisions to make. Is it a 2D game? Maybe use something like Pygame, Löve (for Lua), etc.
At some point you'd be looking into Godot, Unity or similar. And guess what, you could take your basic text-based game, and re-use parts of it as the brains for your game.
Don't get into the game-engine building rabbit hole. It's very fun if you are into that, but know you are unlikely to release anything by going that route. Ask me how I know.
Platform: start with whatever machine you use for development. Presumably you know a lot about it, don't start a side-quest :) Specially, avoid consoles for now. Cross-platform development is getting easier than ever - sometimes all you need to do to start running your game on mobile is to click a dropdown. But that simplicity is deceiving, there's lots you'll have to learn about other platforms. Stick with what you know, until you are comfortable.
I'll let others provide reference material, my sources are outdated as I'm past my Gamedev.net days (for the time being).