A Machine Learning Model for Understanding How Users Value Designs: Applications for Designers and Consumers

Abstract

In this thesis, I demonstrate a number of advances toward developing a machine learning (ML) model of how designs are valued by their users. The model can be used to better understand the implications of furniture design decisions, as well as for commercial strategy.

Existing ML systems have been trained on the physical and aesthetic features of completed furniture designs. We consider these methods to be “top-down” because designers and software engineers alone determine which features are considered important to the value of a design. To better capture the nuances of how users actually value the various functions of their furniture, I first develop a framework for ingesting and classifying user feedback. Next, I conduct a user survey to test this framework, generating a “bottom-up”, labeled dataset from the feedback, requiring no post-processing. Finally, I develop methods for the computational analysis of this data. The analysis is based on a probabilistic ML model trained on the real user data collected. The model is trained to quantify how users value various features of furniture designs, beyond only physical and aesthetic features. I show how the model can augment existing datasets and produce data visualizations to inform design practice and commerce.

This framework represents a step toward a future in which data sets for furniture—and other design domains—are more accessible. By making user feedback available to designers at scale, and establishing methods for collecting this data, we can accelerate the development of designer intuition and deliver significantly greater value to more users.