Data analytics in transport with emphasis on railways

• Formally define problems and identify the relevant analysis
• Choose, collect, format, and clean data according to recommended standards
• Describe data by means of descriptive statistics, ANOVA, and test for significance
• Explain the theoretical basis for linear regression and “ordinary least squares”
• Perform residual analysis, apply data transformations, and evaluate model fit
• Fit multiple regression models with stepwise regression
• Formulate decision models with logistic regression
• Formulate demand and failure models with Poisson regression
• Describe operations problems in railways and applications of data analysis in planning and forecasting
• Evaluate case studies in forecasting or causal analysis
• Develop skills in relevant statistical analysis software, especially “R”

The course introduces linear regression and then continues to advanced methods. Learning objectives cover both methods and problem solving tips. A variety of data sets from mobility and transport are explored to demonstrate complex models, but also to demonstrate what can go wrong, and how to avoid false interpretations. The course ends with an emphasis on real world data and the possibility of multiple interpretations of that data.

The course meets together with diplom course 62637, but the bachelor students receive different exercises and evaluation.

62668, Students should have previously passed a course covering descriptive statistics, distribution fitting, and hypothesis testing. This is a firm requirement.

Lectures, exercises, some group work and final exam.

Research group: Production, Transportation and Planning
Elective: Decision Support in Engineering: 3/5. semester