In this study, we measure velocity variations during two cycles of crustal inflation and deflation in 2020 on the Reykjanes peninsula (SW Iceland) by applying coda wave interferometry to ambient noise recorded by distributed dynamic strain sensing (also called DAS). We present a new workflow based on spatial stacking of raw data prior to cross-correlation which substantially improves the spatial coherency and the time resolution of measurements. Using this approach, a strong correlation between velocity changes and ground deformation (in the vertical and horizontal direction) is revealed. Our findings may be related to the infiltration of volcanic fluids at shallow depths, even though the concurrent presence of various processes complicates the reliable attribution of observations to specific geological phenomena. Our work demonstrates how the spatial resolution of DAS can be exploited to enhance existing methodologies and overcome limitations inherent in conventional seismological datasets.