# Amazon Summer Course in Data Science and Quantitative Economics

## Summary

In keeping with its support for diversity in science and its ambition to become the world’s best employer, Amazon is investing globally in the next generation of outstanding young programmers and researchers. As part of this process, beginning in 2022, Amazon is co-organizing and sponsoring an online Summer Course in Data Science and Quantitative Economics in partnership with [QuantEcon](https://quantecon.org), the [African School of Economics](https://africanschoolofeconomics.com/), and the [ENSEA](https://ensea.ed.ci/en/). The target audience for the course is young economists from underrepresented communities who aspire to a career in data science or computer-driven economic modeling. The goal is to provide talented students from diverse backgrounds with an introduction and exposure to key tools in quantitative economics, which are increasingly in demand in the rapidly transforming economics job market.

The lead instructor and academic coordinator for the program is [Thomas J. Sargent](http://www.tomsargent.com/), Professor of Economics at New York Univeristy and 2011 Nobel Laureate in Economic Sciences.

## Dates
* [to be added]

## Syllabus

This series of 10 sessions will cover topics from the following list:

1. foundations of programming in Python
2. data engineering with Pandas
3. data visualization in economics and social science
4. Bayesian and frequentist probability
5. information
6. Linear projections and their uses
7. singular value matrix decompositions for econometrics and statistics
- application to principal components analysis (PCA)
- application to dynamic mode decomposition (DMD)
- covariance decompositions
- Cauchy-Schwarz inequality and $R^2$

8. Markov chains and applications
9. dynamic programming
- theory
- algorithms
10. linear optimal control theory
11. optimal filtering and smoothing
- duality between filtering and control
- the Kalman filter
- recursive representation of likelihood functions
12. rational expectations models
- equilibrium concept
- econometric implications and Lucas critique
13. Search and matching models
- the lake model
- the McCall model and extensions
14. asset pricing
- the Hansen-Richard approach to implications of $E Rm =1$
- the Harrison-Kreps model of bubbles
16. Bayesian and frequentist statistics
17. estimation and prediction theory
18. income and wealth distributions
- dynamics
- invariant distributions
- cointegration and scaling
19. Ramsey and Stackelberg problems
- linear-quadratic examples
- micro examples
- macro examples
21. economic and social networks
- applications to input-output models
- applications to social learning
- contagion in financial networks
22. linear programming
- optimal transport
- Leontief and von Neumann models
- mechanism design
23. auctions
- first- and second-price auctions
- introduction to simultaneous sales of multiple goods
25. Economic history
- application to fiscal-monetary histories
- application to costs of government finance
27. introduction to high performance computing

## Prospective Organization

Classes will meet for 3 hours, with a break of 15 or 20 minutes in the middle.

Each class will aim to have about half "theory" and have "applications", as the topic under study permits.

We shall provide students ways to apply what they learn
in each class.

1. Classes will gradually build on earlier classes.

1. Students are encouraged to form study groups to foster working in teams (a valuable skill in both industry and government and academia)

2. Before each class, study groups will be provided materials to help get ready for the next class and also help solidify understandings of earlier classes.

3. We hope that study groups can include students with mixtures of skills and mathematical and statistics backgrounds. (This also typically happens in industry and government and academia.)

4. Each class will end with assignments and recommendations of things to be studied before the next class.

## Learning Outcomes

Participants will learn how to connect economic modeling problems to numerical implementations in Python using clear, efficient and effective code. Participants will be able to apply basic software engineering principles to organize, share and collaborate on coding problems.

These skills are in high demand for economists and data scientists at Amazon and similar organizations around the world.

## Student Evaluation

* Homeworks that combine math and social science content with Python programming after each class
* Formation of teams to scrape data, reduce its dimension, and interpret it
* End of program "project" to be presented to the entire class

## Teacher Evaluations

* Avenues for students to criticize content and delivery after each class

## Eligibility and Registration

Ideal candidates will already have had some exposure to probability, linear algebra and principles of economics, up to the level of a high-quality undergraduate economics or computer science program.

## Internship Program

At the end of the course, participants will be evaluated on the technical skills acquired. Internship opportunities are available at Amazon for the most outstanding candidates.