Seasonal Green view index（SGVI-2023） vision2

The SGVI-2023 v2 dataset represents an updated version of the original SGVI-2023 dataset, which is publicly available at 10.6084/m9.figshare.28869701. In this revised version, missing street-level Green View Index (GVI) values in the original dataset have been systematically completed.To address data gaps caused by the unavailability of Sentinel-2 imagery, an optimal predictive model was developed using Sentinel-1 data and quarterly Planet remote sensing imagery as inputs. This model achieved strong performance at the point scale, with a correlation coefficient (R) of 0.83084, a mean absolute error (MAE) of 0.04272, and a root mean square error (RMSE) of 0.06082.For missing values resulting from the absence of Sentinel-1 data, a separate optimal model was constructed using Sentinel-2 imagery, quarterly Planet data, and the Normalized Difference Vegetation Index (NDVI) as input features. This model demonstrated improved predictive accuracy, with R = 0.8789, MAE = 0.03609, and RMSE = 0.05213 at the point scale.The two optimized models were subsequently applied to fill the corresponding missing values in the original dataset. Through this process, a more spatially complete and internally consistent street-level GVI dataset was generated.