Data science and data visualization

This chapter focuses on a select number of effective multidimensional techniques and their application in the transportation domain. Generating an effective multivariate visualization is more challenging and usually involves more compromises than visualizing two or three variables. This section focuses on interactive visualization techniques that can be used to explore the data and uncover clusters and relationships. Parallel coordinates, t-distributed stochastic neighbor embedding (t-SNE), and multidimensional scaling (MDS) are applied on a number of publicly available transportation data sets. It is shown that MDS and t-SNE provide an effective way to visualize dozens or hundreds of variables through dimensionality reduction.