We show that subspace clustering on sequential data can be dramatically improved by incorporating spatial constraints. View on Google Scholar.