Data Analytics in Environmental Health and Engineering

3 Credits.

Data analytics is a field of study involving computational statistics, data mining and machinelearning, to explore data sets, explain phenomena and build predictive models. The coursebegins with an overview of some traditional analysis approaches including ordinary leastsquares regression and related topics, notably diagnostic testing, detection of outliers andmethods to impute missing data. More recent developments are presented, including ridgeregression. Generalized linear models follow, emphasizing logistic regression and includingmodels for polytomous data. Variable subsetting is addressed through stepwise procedures andthe LASSO. Supervised machine learning topics include the basic concepts of boosting andbagging and several techniques: Decision Trees, Classification and Regression Trees, RandomForests, Conditional Random Forests, Adaptive Boosting, Support Vector Machines and NeuralNetworks. Unsupervised machine learning approaches are addressed through applications usingk-means Clustering, Partitioning Around Medoids and Association Rule Mining. Methods forassessing model predictive performance are introduced including Confusion Matrices, k-foldCross-Validation and Receiver Operating Characteristic Curves. Public health andenvironmental applications are emphasized, with modeling techniques and analysis toolsimplemented in R. EN.570 616 meets with EN.570.416 . Undergraduate (usually Senior) students should sign up for 416 with permission of instructor only.