Advanced Data Analytics

Course introduction

This course will run in parallel and cover more advanced methods of data analysis for policy evaluation, building upon the core period course ‘Urban Data Analytics’. Students will be introduced to non-linear regression models, panel data analysis, binary dependent variable, and instrumental variable regression. A major focus of this module lies on the interpretation of results with the aim to conduct evidence-based policy analysis.  
 
Weekly lectures and hands-on workshops will introduce the student to both theory and practice. During workshops, students will work with datasets using the statistical software Stata to translate knowledge into practice. Moreover, as the course evolves, students will recognize how different econometric methods are applied within innovation, resilience, and regional economic development literature. The objective of Module 2 is to gain an advanced understanding of quantitative data analysis methods for urban economic analysis.  
 

Course objectives

The overall aim of the course is to gain insights into the fundamental urban and regional economic processes that explain economic development and resilience, and to develop the ability to analyse and govern these processes in practice.  
 
More specifically we defined the following objectives:

• Reproduce and interpret theories that deal with urban and regional economic development and resilience.  
• Apply theories, concepts and analytical research methods from contemporary literature to practical examples. 
• Apply economic and geographical reasoning to analyse economic development and governance on various spatial scales: from countries to regions to neighbourhoods to households and individuals (micro-level, meso-level and macro-level analyses). 
• Identify criteria for urban and regional economic analysis using quantitative data analysis methods for evaluation. 

Course content

Week 1: Recap OLS
In this workshop we will review some of the key concepts concerning OLS modelling. Students must prepare for the workshop based on a 10-question guide.
Week 2: Non-Linear Regressions
We will explore why modeling non-linearities may (or may not) make sense, polynomial regression models, logarithmic regression models and interaction terms.
Week 3: Non-Linear Regressions (cont’d) 
The workshop will focus on binary and continuous interaction terms.The tutorial preparation requires studying the key concepts and analysing econometric output discussed in Chapter 8 from Stock & Watson.
Week 4: Panel Data Analysis 
The lecture will introduce students to panel data models and their application. We will focus primarily on fixed effects models based on Chapter 10 from Stock & Watson.
Week 5: Binary Dependent Variable 
Students will be introduced to different modelling techniques when the dependent variable is binary, i.e. it can only take two values, 0 and 1. This includes linear models, i.e. LPM, and two non-linear models, i.e. logit and probit. 
Week 6: Instrumental Variable Regression 
Students will be introduced to a new modelling technique, instrumental variables. This is econometric approach is used to address issues stemming from circular causality, that is when causality runs both from X to Y and from Y to X. 
Week 7: Q&A Session 

Course information