Data scientists use computer models to semi-automatically classify deprived urban areas from satellite imagery and other spatial datasets, which allows mapping efforts to be scaled up. Developments in deep learning methods show that well-trained models can achieve accuracy of more than 90%. However, such methods require a large amount of high-quality training data, expensive very high-resolution imagery, and they are computationally demanding. Consequently, most computer models of urban deprivation typically cover small study areas within a single city.

In practice, the data used to model deprived areas overwhelmingly represents physical characteristics such as building size, shape, and orientation. Very few models consider social characteristics such as presence of trash piles, crime rates, or zoning designations. This is because data about social services and infrastructure are sometimes not well synthesized across neighborhoods and cities, and/or must be gathered from multiple sources. As a result, computer models of urban deprivation mainly reflect informal settlements, and are less useful in contexts where the urban poorest live in durable housing but face multiple social deprivations.

Furthermore, a majority of image classification models result in maps with discrete boundaries between area types, however, deprived areas may not have sharp boundaries.