ResiDense estimates the appropriate development density in units per hectare for a given site using a machine learning regression model1 trained on 10 years of London planning consent data.
The model looks at the relationship between development density and key planning constraints and designations, alongside a variety of geography-related site properties such as Public Transport Accessibility Level (PTAL) and local population density2. Estimates are made based on these site properties and are then adjusted according to actual nearby consents.
ResiDense also identifies relevant place-specific policy designations, finds planning consents of a similar density and size and identifies affordable housing and unit mix requirements. Policies and requirements are derived from policy maps and documents published by local authorities.
ResiDense is an independent digital planning project created by Aman Sahota. A density estimate from ResiDense is for information only and is not a guarantee of its viability or policy compliance.
1 The regression model is trained and run using the TensorFlow Decision Forests library.
2 The data used to train the model is obtained from the Greater London Authority, Transport for London, the Office for National Statistics, the Department for Levelling Up, Housing and Communities, and Emu Analytics.
Copyright © 2024 Aman Sahota