Introduction: The GLOW-S dataset provides a comprehensive record of global river widths, captured with high-resolution Sentinel-2 imagery. This dataset focuses on delivering an enhanced spatial and temporal analysis of river widths, suitable for diverse hydrological and environmental studies. Data Collection and Processing: Image Source: River width data were extracted from Sentinel-2 satellite imagery accessed via Google Earth Engine, utilizing higher resolution capabilities than previous datasets based on Landsat. Extraction Methodology: Following validated methods from prior research (Feng et al., 2019; 2022), the dataset involves: Identification of rivers observable in Sentinel-2 images. Construction of cross-sections for width extraction. Application of quality filters on images to ensure accuracy in width measurements. Determination of water extent and subsequent river width calculation at each cross-section for images spanning 2017 to 2022. Dataset Features: Temporal Coverage: Data spans from 2017 to 2022, capturing seasonal and interannual variations in river width. Spatial Resolution: Utilizes the superior spatial resolution of 10m from Sentinel-2 for detailed river width analysis. Postprocessing: When multiple observations for the same cross-section and date exist, the dataset averages these widths to enhance usability. Column names: riverID (cross section identifier): RXS date (image capture date): YYYY-MM-DD width (river width captured in meters) Additional data: The shapefiles of cross sections are also attached alongside.