Course Description

“Advanced” research strategies are, by virtue of their name, often complicated and require full courses of their own. Just look at the range and specificity of courses available via online learning platforms, the volumes of tutorials on building complex machine learning models, or appropriately estimating causal effects. Too often, however, analysts (and their clients) can place huge emphasis on the apparent benefits of this complexity without pausing to consider a method’s appropriateness, or limitations inherent to the data itself that may scupper valid analysis.

This course is designed to focus on the fundamentals of advanced data analytics, and in doing so to draw attention to the various theoretical, inferential, and computational issues that ultimately constrain any quantitative project. Most importantly, this course will equip you with the skills necessary to know when and why certain procedures are useful, and how to ensure the validity and appropriateness of your analysis. Along the way, we will cover a range of exciting and cutting-edge techniques across several key areas of analysis, including causal inference, machine learning, and experimentation.

This course should refine your ability to conduct advanced analysis of data across a variety of contexts, and I hope it sparks further interest in the various methods you will encounter along the way. Of course, there is only so much material eight modules can cover and, where applicable, I have highlighted extra readings you may wish to pursue beyond the confines of this course.

This course is comprised of eight, self-guided modules, two problem sets, and a final test:

1. Effective Analysis in an Era of Advanced Data Analytics
2. Data Visualization and Story-telling
3. Fundamentals of Causal Inference
4. Experimental Survey Methods for Policy Evaluation

 Problem Set 1:

5. Fundamentals of Machine Learning
6. Applying Machine Learning in a Causal Inference Setting
7. Data Integrity and Handling Missing Data
8. Multi-Level Regression and Post-Stratification

Problem Set 2:

Final Test (multiple choice)

Course Objectives:

This course aims to improve your ability to analyze data by focusing on advanced topics in data science and statistics. By the end of this course, you should be able to:

• Recognize appropriate statistical uses for different types of data
• Understand how to translate data into effective visualizations
• Distinguish between correlation and causation, identification and prediction
• Understand the benefits and limitations of machine learning
• Resolve common data issues using advanced statistical techniques

Important Notes:

1) Participants who register to take this course and pass the final examination with a score of 70% or higher will be eligible to receive the APO certificate. Please note that the final examination can be taken only once. Therefore, you decide when you choose to take the examination. Please note that the self-assessment quizzes are for your own evaluation and have no connection with the final examination results.

2) Participants who perform well in this course and receive the APO certificate will be given preference, on a merit basis, for selection to attend the follow-up face-to-face multicountry APO project, provided their nominations are received through the concerned NPOs and slots are available.

3) Notes 1 and 2 are applicable only to participants from APO member countries. Those from nonmember countries are welcome to take the course for self-improvement, although they are not eligible to attend follow-up face-to-face multicountry APO projects.