Methods Lectures

Modern Trends in Data Mining

Trevor Hastie
Professor, Department of Statistics & Department of Health, Research and Policy, Stanford University

Abstract: As their ability to capture and organize large amounts of data increases, organizations rely more on data mining technology to learn from this valuable resource. We will give several examples of this process, based on our own experiences.  This lecture will give a brief overview of some of the most promising new methods for “supervised” learning, including the lasso, random forests, boosting, and support vector machines.

Internet Surveys

Douglas Rivers
Professor, Department of Political Science, Senior Fellow, Hoover Institution, Stanford University

Abstract: This lecture will provide an overview of recent developments in internet surveys.  It will consider the advantages and disadvantages of internet surveys relative to other modes (such as phone, mail, and face-to-face). Recent developments for sample selection, matching, weighting and modeling data with non-response and self-selection will be described, along with applications to elections, public policy and consumer choice.

Bayesian Data Analysis

Simon Jackman
Professor, Department of Political Science, Stanford University

Abstract: Bayesian data analysis uses Bayes’ Theorem to update beliefs about parameters and hypotheses based on data. The lecture reviews conceptual distinctions between Bayesian and (conventional) frequentist inference, emphasizing the conceptual simplicity of the Bayesian approach. Practical examples from social science also demonstrate the utility of Bayesian inference.

Spatial Statistics and GIS

Patricia Carbajales
Geospatial Manager, Earth Sciences Library, Stanford University

Abstract: This lecture will provide an overview to spatial statistics and “geographic information systems” (GIS).  It will discuss the basics of GIS software, data models, shapefiles, geodatabases, earth reference systems, projections, visualization, and spatial analysis.

Bureau of Justice Statistics Data

Thomas H. Cohen
Statistician, Bureau of Justice Statistics

Abstract: This lecture will provide an overview of core data collection activities related to state and federal courts by the Bureau of Justice Statistics.  The lecture will cover several topics, including: (1) the processing and sentencing of felony defendants in state courts; (2) data on civil litigation in state court systems; (3) the organizational characteristics of courts, prosecutors, and indigent defenders; (4) the disposition of criminal appeals; and (5) the Federal Justice Statistics Program linked data file.