101 - 110 of 547 results for: Medicine

A sea change has occurred in science, technology, medicine, politics, and society as a whole: many of the world's biggest discoveries and decisions are now being made on the basis of analyzing massive data sets, referred to as "big data". Everyday examples include social-network friend recommendations, and weather predictions far more accurate than a decade ago; both use vast collections of data to model the past and predict the future. But it is surprisingly easy to come to false conclusions from data analysis alone. For example, we might conclude that acne medicine prevents heart attacks and strokes, if we forget to factor in the age of the patients. Privacy is a major concern: Target stores analyzed buying patterns to predict with remarkable accuracy which of their shoppers had just become pregnant, but trouble arose when they sent baby ads to the homes of pregnant teens whose parents weren't yet in the know. We will start by surveying the history of data-driven activities, leading up to the recent Big Data explosion. A variety of data analysis techniques will be covered, leading students to appreciate that even simple techniques can go a long way when the data set is large enough. Common stumbling blocks leading to false conclusions will be discussed, and students will be asked to debate the many issues surrounding privacy. In one project, students will see whose analysis techniques can best predict user movie ratings based on past rating behavior. A second project will be individually designed in an area of the student's choosing. The seminar will include a mix of assigned readings, small-scale investigations and assignments, classroom discussions, and two projects. No computer programming or special math skills are required; students will learn the basic techniques and tools they need to complete the data analysis assignments and projects.