Timely Data Analysis: Modeling and Exploiting Temporal Correlation with Dr. Nathan Wycoff
Abstract: Making management decisions in business or policy decisions in government often involve taking into account the future state of some aspect of the world. Decision makers have long sought to use present day trends to anticipate what will happen next. In this workshop, we will start by covering the basic statistical tools often used by forecasters through a series of case studies. We will start with basic time series models such as ARIMA models before covering some of the latest machine learning tools based on neural networks that have been developed for the purposes of extrapolation into the future. Throughout the workshop, we will discuss digital data sources which can be useful in developing forecasts. We’ll conclude with a discussion of the prospects and limitations of forecasting. Workshop activities will be conducted in Python.