The CARTO Analytics Toolbox has extended the geospatial capabilities of Google BigQuery with spatial SQL, unlocking advanced spatial functions such as data transformations and clustering. The toolbox now includes Geographically Weighted Regression (GWR) method, a statistical regression technique that models local relationships between variables and an outcome of interest. GWR allows for detecting spatially-varying relationships between the variable of interest and relevant attributes, which can reveal local differences and contrasting relationships in different parts of a study area. The toolbox's implementation of GWR takes advantage of spatial indices to construct neighborhood matrices, allowing users to specify kernel types such as uniform, triangular, quadratic, and Gaussian. By applying GWR to real-world data, users can gain insights into the associations between variables at a local level, revealing patterns that may be masked by global statistics.