We consider the combined use of visualization and sound for uncovering important structure in high-dimensional data. Our approach is based on Markov chain Monte Carlo (McMC) simulations. McMC is a popular computational tool for making analytical inferences from complex, high-dimensional probability densities. Given a particular target density p, we simulate a Markov chain that has

p as its stationary distribution. We propose a new tool for

exploratory data analysis based on an audio representation of McMC output. Several audio streams provide us with information about both the behavior of the Markov chain as well as characteristics of the target density p. We apply this method to the task of identifying structures in high-dimensional data sets by taking p to be a nonparametric density estimate.

In this paper, we present a detailed description of our sonification design and illustrate its performance on test cases consisting of both synthetic and real-world data sets. Sound examples are also given.