Cities experience the urban heat island (UHI), which continue to pose challenges for humanity's increasingly urban population. Past research has revealed that land cover composition and configuration, along with other geographical phenomena (i.e., albedo), can explain much of the spatial pattern of UHI, yet advances await. In response, this research was made to: (i) assess the spatial pattern of mean ambient night temperature across 34 streetscapes in New York City (NYC); (ii) create and differentiate global and local regression models between-natural and built streetscape characteristics- and mean ambient night temperature; and (iii) use geographically weighted regression (GWR) to assess local patterns of correlated associations. Urban canopy layer (UCL) temperatures were recorded across 34 weather stations, and landscape metrics calculated from 0.914 m land cover data with 96% accuracy. Local Getis-Ord Gi* statistic exhibited significant spatial cold and hot spots of UHI in NYC. Global inferential tests revealed that sky-view factor, photosynthesis activity, elevation, and road configuration were the strongest predictors of mean ambient night temperature. Six multiple regression models were ultimately made with GWR fitting the UHI aptly (R2 = 65–74%). Important explanatory covariates were illustrated using local pseudo-t statistics and linked to mean ambient night temperature, supporting the importance of GWR for understanding local UHI interactions. Results also confirm that landscape configuration metrics are stronger predictors of UHI than composition measures. Streetscape design, particularly road patterns and process, requires more consideration when attempting to mitigate UHI during future sustainability planning, urban renewal projects, and research.