We really shouldn't have tried to develop a rhythm game for a game jam without any previous experience.
We developed our game in Unity because that's where we are most experienced in (Not experienced enough, as it turns out)
There was one main problem to solve: Synchronizing stuff to the beat.
At first we simply tried to use the current audio position as a timer and we thought it worked relatively well. Later (15 hours before the end!) we noticed massive accuracy problems. Maybe Unity's update function is just to slow, or idk. We ended up switching to the fmod API (8 hours left) and used tempo markers for timing. We had to do some pretty scary low-level calls none of us had much experience with. This solved the accuracy issues but also cost a lot of time.
The next problem was to check if the keystrokes are really on the beat (or in a time window around it). The actual mathematical problem was not that hard to solve, but we were already very tired and therefore (again) wasted many hours on it. We ended up using a moving average of the time between beats to calculate when the next beat would happen, so we could check if we hit/miss/skip the last OR the next beat.
We got the rhythm part of the gameplay working with only about 3 hours to go and there are still many features/gameplay elements missing but no time (or motivation) left.
TL;DR: Don't try to create a game for a jam that exceeds your skill set by THAT much.
If you want to try our game anyways click here at least it has pretty water shaders