Big data methods in the social sciences

- Supervised methods: models that predict group membership (e.g., gender) or continuous outcomes (e.g., job performance ratings).
- Unsupervised methods: models that determine the appropriate number and nature of clusters (e.g., types of managers) or associations in the data (e.g., consumer preferences) without any other external or objective information.
- Cross-validation methods: a set of methods to ensure that models can fit to fresh data (data that did not participate in model development).
- R and RStudio: R is a free and popular tool for running big data packages and writing programs; RStudio is a free popular interface for making the best use of R.
- Python, Pandas, and SciPy: Python is a free and popular programming language for analyzing big data, and the Pandas and SciPy libraries provide free analysis packages.