The past few decades have brought critical challenges in storage and analysis of large datasets. In this talk, I will first briefly survey some important techniques from Applied Harmonic Analysis that address these challenges. Then I will focus on Diffusion Maps, an algorithm in which the eigendecomposition of a graph operator, the Laplacian, is used to efficiently encode important, sometimes hidden, information about datasets. I will show various Diffusion Maps’ applications, including results as crucial as detecting early stages of blindness and some exciting new findings on autism, both via analysis of medical images.