One of the main obstacles in understanding galaxy formation is the unknown physics of star formation and feedback. The reason as to why there is so much mystery behind star formation and feedback is that this process occurs on a wide variety of spatial and temporal scales. In this project we use IllustrisTNG — a large, cosmological hydrodynamical simulation of galaxy formation — to study the physics of the time variability of star formation. We investigate how the star-formation rates of individual galaxies change over time by performing a detailed time-series analysis. Specifically, we compute the temporal power spectral density of the star-formation history to estimate the amount of power on each fluctuation timescales. We do this for simulations with different physical recipes for star formation and feedback. This allows us to study how different physical processes shape the time variability of star formation. Additionally, we make predictions on the time variability of star-formation histories of galaxies, which can be tested with new observations in the future.