Cluster analysis is a fundamental data analysis technique, but many clustering methods have limitations, such as requiring initial starting points and requiring that the number of clusters be specified in advance. This paper describes an evolutionary algorithm based rough clustering algorithm, which is able to overcome these limitations. Rough clusters use sub-clusters called lower and upper approximations. The lower approximation of a rough cluster contains objects that only belong to that cluster, while the upper approximation contains objects that can belong to more than one cluster. The approach therefore allows for multiple cluster membership for data objects. This rough clustering algorithm was tested on a large data set of perceptions of city destination image attributes, and some preliminary results are presented.