The project aims to classify global urbanity levels using a spatial model that aggregates population density, nighttime light data, and road network infrastructure. The model is built at a quadkey grid level of 15, which has cells of approximately 1x1km at the equator. Six different urbanity classes are defined, ranging from remote areas with low population density to very high-density urban areas in metropolitan centers. The method involves transforming input datasets into the quadkey grid specification, pre-processing nighttime light data to remove outliers, and combining population density and light intensity data using Principal Component Analysis. Spatial dependence is included by calculating a Z-score for each cell based on neighboring cells' values. The resulting index is then used to cluster areas into urbanity classes. The model has been applied to 251 countries, with results showing that 42% of the global population resides in at least low-density urban areas, and that Asia and North America have higher percentages of remote areas due to large countries like Russia and Canada.