Thanks for trying it out!
This was our first game jam together (for most of us, our first game ever) and we were really pressed for time, haha. Unfortunately, we didn't get a chance to add as much content to the game environment as we wanted. We planned on adding map obstacles and enemies at the very least.
As for the proc gen, it isn't based on a noise density function--it doesn't control individual tiles at all, actually. We have a collection of preestablished map sectors that are pieced together on a grid, and we use a pathfinding algorithm to verify that it's possible for the player to traverse the map. The bottleneck here, and the reason maps often look similar, is just that we didn't have time to create lots of unique sectors. There are around 20-30 as it stands. From a game design perspective, it was difficult to create additional sectors without ruining the flow of the game. Sectors are balanced around how likely they are to be included in a successful pathfinding traversal; for instance, sectors that typically result in a dead end generally have more coins, fuel, and booster pads, so the player is rewarded for exploring and isn't punished for taking a wrong turn. Booster placement is also a consideration, as we want to encourage smooth movement between sectors by raising the height of ceilings above them.