Architecture and affordance: a data driven and computational approach to the architectural analysis and design of buildings using affordance as a model of typology

Abstract

ISpace syntax has long been used as a spatial analysis tool to graphically represent the relationships and connections between spaces. This has been used to promising effect in the past to compare buildings of a similar typology in order to better understand them. We can use the framework of space syntax and introduce affordances to gain a better understanding of how affordances of access, natural light, sound, and activity congregate in building typology. By building a digital sensor array and conducting an analysis of a particular building typology, we can start to find optimal patterns of affordance which exist within living buildings. The typology I am looking at for this thesis is the United Churches of Sudbury. Sacred space is a phenomenologically dense and interesting typology which lends itself to generating interesting data. United Churches as a typology also have a philosophy of shared multi-use space which lends itself well to this more generalized approach to understanding program through affordances. In this thesis, I look at four different case studies of United Churches, one to understand what makes a United Church, and three others as examples of typology and subjects for collecting data. The aim of this work is to use this data to develop a set of computational design tools that will eventually be used to suggest a possible design for a United Church on the site of Larchwood Memorial United Church in Dowling. Aside from site analysis, there are two parts to this affordance based process. Just like how Gibson distinguishes between affordances and the invariant properties of objects, I take stock of affordances through a set of affordance graphs and tables of relevant data from each of my case studies in Sudbury. I also look at the design solutions such as furnishings and building openings which these churches used to satisfy those affordances and document them in the from of a pattern language. Using the affordance data, I find common patterns of design and layout for a given typology. These patterns of affordance can then be used with a generative algorithm to generate a schematic design. In this thesis, I will present a number of schematic designs and explore the range of outcomes, limitations, and areas for improvement that you can expect from this approach to generative design. A final schematic design can then be matched with examples of vernacular objects and strategies found and documented through the pattern language of sacred space. Using data that I have collected about openings and furnishings in the case studies, it becomes possible to automatically populate these schematic designs with objects to complete the design. This thesis will detail this process and the theory behind it.