Predictive analytics comprise a variety of techniques that predict future outcomes based on historical and current data. In practice, predictive analytics can be applied to almost all disciplines from predicting the failure of jet engines based on the stream of data from several thousand sensors, to predicting customers’ next moves based on what they buy, when they buy, and even what they say on social media. At its core, predictive analytics seek to uncover patterns and capture relationships in data. Predictive analytics techniques are subdivided into two groups. Some techniques, such as moving averages, attempt to discover the historical patterns in the outcome variable(s) and extrapolate them to the future. Others, such as linear regression, aim to capture the interdependencies between outcome variable(s) and explanatory variables, and exploit them to make predictions. Based on the underlying methodology, techniques can also be categorized into two groups: regression techniques (e.g., multinomial logit models) and machine learning techniques (e.g., neural networks). Another classification is based on the type of outcome variables: techniques such as linear regression address continuous outcome variables (e.g., sale price of houses), while others such as Random Forests are applied to discrete outcome variables (e.g., credit status). Predictive analytics techniques are primarily based on statistical methods. Several factors call for developing new statistical methods for big data. First, conventional statistical methods are rooted in statistical significance: a small sample is obtained from the population and the result is compared with chance to examine the significance of a particular relationship. The conclusion is then generalized to the entire population. In contrast, big data samples are massive and represent the majority of, if not the entire, population. As a result, the notion of statistical significance is not that relevant to big data. Secondly, in terms of computational efficiency, many conventional methods for small samples do not scale up to big data. The third factor corresponds to the distinctive features inherent in big data: heterogeneity, noise accumulation, spurious correlations, and incidental endogeneity.