Style Analysis Methodology: Identifying the Car Brand Design Trends through Hierarchical Clustering

Abstract

This paper aims to identify car design trends within various automobile manufacturers by investigating two objectives: first, finding similarities between car styles among different car brands from various automobile manufacturers to specify unique car designs which lead the trend; second, identifying the consistency of the brand design characteristics through hierarchical clustering. To do that, Fourier decomposition was used to quantify the car design similarities between 120 cars from 23 different brands. The calculated similarity index is then compared with network centrality measures to identify the clustering of the car brands. The quantified style data then can be applied to accurately predict the design trend. Thus this study can contribute to identify car style trends for strategic design decisions.