The U.S. Postal Service has been using machine learning and scanning technologies since 1999. Because its postal offices have to look at roughly half a billion pieces of mail every day, they have done extensive research and developed very efficient algorithms for reading and understanding addresses. And not only the post office, ATMs can recognize handwritten bank checks. Evernote can recognize handwritten task lists and Expensify can recognize handwritten receipts. In this project, I used K-means clustering (the algorithm behind this magic) and scikit-learn to cluster images of handwritten digits