SummaryCalcium imaging has become an increasingly popular way to probe the activity of astrocytes. However, the governing principles of astrocyte calcium dynamics are still elusive and their relationship to cellular events ill-defined. Useful assumptions and ‘shortcuts’ commonly applied to neuronal recordings therefore do not hold true for astrocytes. The imaging of astrocyte calcium activity per se can be relatively straightforward, subsequent analysis methods that adequately capture the richness and complexity of calcium dynamics remain scant. Here, we introduce STARDUST, a pipeline and python-based data processing for the Spatio-Temporal Analysis of Regional Dynamics & Unbiased Sorting of Transients from astrocyte calcium recordings. STARDUST builds upon AQuA to identify patches of active signals, from which it builds a data-driven map of regions of activity (ROAs) that can be combined with cell-segmentation and/or correlated to cellular morphology. For each ROA, STARDUST extracts fluorescence time-series, and performs signal identification and features extraction. STARDUST is agnostic to cell morphology (or cells altogether) and putative calcium propagation across ROAs. Instead, it focuses on decomposing calcium dynamics in a regionalized fashion by treating ROAs as independent units, for instance allowing investigations by signal feature-based ROA rank. STARDUST also identifies ROAs as “stable” (active throughout), “ON” (turned on during drug application) and “OFF” (turned off during drug application) in pharmacology experiments, permitting studies of astrocyte calcium “micro-domains” based on their functional responses. With a systematic set of instructions and troubleshooting tips, and minimal computational/coding background required, STARDUST is a user-friendly addition to the growing toolbox for the exploration of astrocyte calcium dynamics.