[en] In rapid neutron capture, or r-process, nucleosynthesis, heavy elements are built up via a sequence of neutron captures and β decays that involves thousands of nuclei far from stability. Though we understand the basics of how the r process proceeds, its astrophysical site is still not conclusively known. The nuclear network simulations we use to test potential astrophysical scenarios require nuclear physics data (masses, β decay lifetimes, neutron capture rates, fission probabilities) for all of the nuclei on the neutron-rich side of the nuclear chart, from the valley of stability to the neutron drip line. Here we discuss recent sensitivity studies that aim to determine which individual pieces of nuclear data are the most crucial for r-process calculations. We consider 3 types of astrophysical scenarios: a traditional hot r process, a cold r process in which the temperature and density drop rapidly, and a neutron star merger trajectory. The resulting sensitivities highlight the differences in how the r process proceeds in each case. However, there is one key similarity: the most important pieces of nuclear data in all cases are for nuclei near the closed shells and in the rare earth region 10-20 neutrons from stability. We recommend these nuclei be targets of the next generation of experimental campaigns