We give a visual introduction to several geometric techniques for analyzing data. These include both unsupervised learning (clustering, dimensionality reduction, topic modeling), and supervised learning (k-nearest neighbors, support vector machines), though we don't expect you to know what any of those words mean! The goal is to distill the methods down to visual and oral description without mathematical notation. The performance of data analysis techniques will be illustrated on real-world image and text datasets. Mini-course participants will be encouraged to develop their own purely visual explanations