I think you about to do it all starting from wrong end.
First think what you want to get from making that game, ie. indie dev and making real game that sells, making some game as portfolio project (so you can get hired in future), or making it to learn.
From your post i would suggest you make game that you can learn something new from.
So it should be small game (really small), and it should be done and finished in 100%.
L1z4rD89 suggestion was right on spot.
Google for retro games, pick one that is simple to create (no characters animations that maybe for next one)
some more retro games sugestions:
- moon patrol
- asteroids
- lunar landere
- space invaders
- tetris (you can make it 3d now)
- arcanoid
Point is to make new step and finish and publish finished game. So you know how to publish to steam, ios and android.
Doing gameplay that works is about 30-50% of whole code, then fixing bugs, polishing adding for eg. steam support that is like 80% of work for game when coding vertical slice is that 20%.