Applied Statistical Methods

Dec 2018

David Rios possibly has many strengths, none of which include clarity or teaching style. He was tardy to class, dismissed us early, and had dry lectures that were impossible to stay engaged with. His notes make little sense, as he jumps from shortcut to shortcut--he almost seems to take pleasure in being roundabout and offering no real explanations. His problem sets were not as challenging as they were absolutely impossible to decipher the questions he was asking. Rios is obviously smart and has talents, but these are far outside the realm of teaching.

May 2015

Incredible professor! I have taken quite a few classes in the statistics department and I have never had such a great professor (not just in the stat department, but also at columbia in general). He is so dedicated to teaching the class and each of student individually. He clearly puts so much time into preparing for each lecture, and is also extremely helpful when it comes to answering questions. He is very helpful during his office hours and is also open to making time outside his office hours to meet with students. The statistics department could really use more professors like this one!

May 2013

Professor Madigan is an incredible professor. He is extremely receptive to emails and is always more than happy to talk to his students. He is one of the nicest professors at Columbia and is genuinely interested in having his students learn something rather than having them just care about the grade. The man is an amazing lecturer and is able to make difficult concepts understandable. If you put the time into this class, you will walk away with a solid understanding on how to use certain statistical tools. Regression, in a nutshell, is an art form. There are no straightforward, procedural methods, and this class teaches you a variety of statistical tools to put in your statistical toolbox. Although the class begins with some easier concepts (especially if you have taken the prerequisite 2024 Applied Linear Regression Analysis), it gets more interesting in a hurry when he begins discussing generalized additive models. The one challenging portion of this class is the math behind some of these statistical concepts. Luckily, you don't have to fully understand them, but they're still very interesting. There is one homework assigned per week, which requires the use of R. If you didn't know R before this course, you definitely will be comfortable with it by the end. The homeworks, which use the concepts taught in that week's lecture, require a bit of tinkering. Luckily, the TA and the professor are both extremely helpful if you ask. In addition, the textbook is very helpful in completing the homeworks (which is available online). There is a final but no midterm. The final is open book and last about an hour (most people finished in around 30-40 minutes). There are six short response questions that don't test R knowledge, but rather theory behind the things taught in class. It's the kind of final where you either know it or you don't. A bit difficult to prepare for, but as long as you attend lectures and take notes you should be fine. Lastly, Professor Madigan had each student do two projects, which could be done in groups of 2-3. These projects were incredibly enjoyable; students were given the opportunity to research any topic they pleased (I recall one student writing a project on basketball statistics); I recommend you choosing a topic that interests you. You were required to use statistical methods taught in class, like poisson regression or time series. A write-up is also required in the project in explaining your results. Don't leave the projects to the last minute though, as they take a bit of time because R is bound to give you some fits along the way. Really make it something yours and if you put your heart into it (cheesy i know), you will be rewarded.