A new method was developed to map urban areas using a machine learning algorithm that groups buildings based on their density. By analyzing building heights, researchers created a more accurate measure of urban size, including vertical space. They also identified employment centers within these areas. The results show that the new urban areas closely match commuting patterns, are more precisely measured than administrative boundaries, and follow Zipf's law in terms of population, surface area, and vertical space. Additionally, improvements in transportation have a significant impact on the size of urban areas.