Abstract:Fractals are geometric objects that describe real life systems more closely than idealized Euclidean objects. In particular, they provide a simple yet powerful tool to study the non-Gaussian properties of random processes. In this talk, I will introduce fractals from an astronomer's point of view, i.e., through examples and general properties. Then, I will describe a method I developed for analyzing fractal trails. Finally, I will show a simple application of this technique in stellar dynamics, where we observe a non-Gaussian random walk. For this process, fractal analysis not only reveals the non-Gaussian feature, but also gives the relevant time scales.