It is recognised now that a variety of real-life phenomena ranging from diffuson of cold atoms to motion of humans exhibit dispersal faster than normal diffusion. Lévy walks is a model that excelled in describing such superdiffusive behaviors albeit in one dimension. Here we show that, in contrast to standard random walks, the microscopic geometry of planar superdiffusive Lévy walks is imprinted in the asymptotic distribution of the walkers. The geometry of the underlying walk can be inferred from trajectories of the walkers by calculating the analogue of the Pearson coefficient. Introduction. The Lévy walk (LW) model [1-3] was developed to describe spreading phenomena that were not fitting the paradigm of Brownian diffusion [4]. Still looking as a random walk, see Fig. 1, but with a very broad distribution of excursions' lengths, the corresponding processes exhibit dispersal faster than in the case of normal diffusion. Conventionally, this difference is quantified with the mean squared displacement (MSD), r 2 (t) ∝ t α , and the regime with α > 1 is called superdiffusion. Examples of such systems range from cold atoms moving in dissipative optical lattices [5] to T cells migrating in the brain tissue [6]. Most of the existing theoretical results, however, were derived for one dimensional LW processes [3]. In contrast, real life phenomena -biological motility (from bacteria [7] to humans [8] and autonomous robots [9,10]), animal foraging [11,12] and search [13] -happen in two dimensions. Somewhat surprisingly, generalizations of the Lévy walks to two dimensions are still virtually unexplored.