Competition Data Analysis

An analysis of the 1,715 competition submissions, created for the Guggenheim Helsinki Now exhibition

The Guggenheim Helsinki Design Competition inspired one of the largest archives of contemporary architectural thought. After the jury had reviewed all entries and announced the six finalists, the Guggenheim invited Finnish architect and theorist Martti Kalliala and American data scientist and taste researcher Hugo Liu to collaborate on an experimental analysis on the submissions as a data set.

Through a pairing of artificial intelligence and subjective interpretation, the resulting map offers an illustration of themes and trends in museum architecture today.

Method 1: Correspondence Analysis

Hugo Liu performed a computational linguistic analysis on the concept description text submitted by each competition participant. The process extracted recurring key words that scored each project in terms of its emotional, social, spatial, and temporal ambitions. Shown here are some exemplary samples of each.

Liu then analyzed the text for common themes that were present in many of the submissions. In an effort to visualize the relationship between the submissions, the projects were then plotted in a single field.

Method 2: Architectural Pattern Recognition

Responding to Liu’s computational analysis, Martti Kalliala developed a new term-based classification system for analyzing the “personality types” of buildings. Sixty young architects helped to evaluate all the submissions according to the term-based system. The submissions were then mapped computationally and interpreted by Kalliala into an experimental reading of the themes of contemporary architecture.