Non-intrusive Semi-automated Modelling of Ordinary Buildings’ Intermediate Floors: Collective Housing in Porto, Portugal

Contemporary ordinary collective housing buildings constitute quantitatively the greater built heritage of many compact cities, and in the near future they will need to be rehabilitated en masse. Their repetitive patterns promote the application of GIS/BIM modelling technology. Thus, this research develops a shape grammar to semi-automatically estimate the layout of the intermediate floors of residential buildings from their external shape. The grammar is based on the geometric capacity of floors to accommodate rooms in contact with the facade and to be divided into several dwellings. Grammar Part I (rooms) uses the topological concept of ‘cells’ to estimate the spaces behind the facades. Grammar Part II (dwellings and stairs) divides the floor into dwellings and locates the common staircase. This shape grammar is programmed in Arcpy (ArcGIS’ Python library) with variations in accordance with its tools, restrictions, and processing performance. Finally, this shape grammar’s script is applied to one hundred collective buildings in Porto, studied in a previous Ph.D. thesis. The results of this sample illustrate that the shape grammar can estimate reality in many cases, revealing an underlying ordinary model with varying standards. Moreover, it offers a relevant analysis of the buildings’ capacities and morpho-typological characteristics, even for irregular geometries. Its general applicability, non-intrusive methods, and en masse processing can be useful in quantitative analyses and preliminary modelling for urban regeneration studies and policies.