Keio University

DEEP(Data-driven Economics and Econometrics Program)

DEEP is a program in the Faculty of Economics that aims to equip students with the data science knowledge and skills necessary to conduct data-driven economic research. DEEP emphasizes not only the study of theoretical systems of data science, but also its practice. Through the process of hands-on programming, organizing and analyzing real-world data, and presenting results in the form of academic papers or products that address societal challenges, students develop a solid foundation for pursuing careers as data scientists.

Curriculum and completion requirements

The DEEP curriculum consists of the following three course categories.

1. Core courses (courses at Hiyoshi, 10 credits or more)

Linear Algebra*, Advanced Linear Algebra, Introduction to Calculus, Calculus*, Statistics I*, Statistics II*, Introduction to Econometrics* (* denotes required subjects)

2. Research courses (courses at Mita, 16 credits or more)

Probability and Statistics a and b, Econometrics a and b, Advanced Econometrics a and b, Time Series Analysis a and b, Bayesian Statistics a and b, Introduction to Artificial Intelligence a and b, Quantitative Macroeconomics a, Economic Geography a and b, AI Industry Studies

3. PBL (Problem-Based Learning) courses

i. PBL coursework in specific areas

Theory and Practice of Token Economies b, Introduction to Data-Driven Finance b, Data Science Consulting

ii. Seminar paper (graduation thesis)

iii. Research project C paper

After obtaining the required credits in 1. and 2. and submitting the final coursework in 3., a certificate of completion is awarded. Guidelines for the final coursework are explained below.

First, during the two years at Hiyoshi, students study mathematics and statistics, the foundations of data science, which are referred to as “core courses.” Knowledge of mathematical fields is essential for a proper understanding of the theoretical underpinnings of data science. In addition, although not a “core course,” taking a related mathematics course such as Introduction to Probability Theory and an information processing course involving programming languages would be very useful for practicing data science at Mita. In order for students to further develop their understanding of data science, it is highly recommended that they attend the training sessions offered by the AI and Advanced Programming Consortium (AIC).

In the following two years at Mita, students take “research courses” to learn advanced data science methods and applications. The theoretical systems students learn here are all necessary to properly formulate questions for social issues to be solved with data science, as well as for correctly using different methods as appropriate for the characteristics of real-world data. In addition, some research courses allow students to learn the practice of data analysis.

At the conclusion of DEEP, students will submit a final piece of coursework produced through the practice of data science. This is done in Problem-Based Learning (PBL) courses. PBL refers to practical learning conducted for the purpose of problem solving. DEEP’s PBL courses also include the creation of apps to solve problems in specific areas, but the final coursework is a research paper, produced as part of the student’s seminar or research project, that examines the validity of theories and policies in a variety of fields using data science methods.

Please contact the DEEP coordinator for information on what research topics are eligible for the DEEP final coursework.

Final coursework guidelines

In order for a seminar graduation thesis or research project thesis to be recognized as final coursework for DEEP, it must fall into one of the following categories.

Category

1. Theoretical research

Research on the mathematical properties of models and methods used in econometrics, statistics, machine learning, etc.

2. Empirical research

Analysis of real-world data using models and methods that are the subject of theoretical research

3. Numerical experiments

Validation of the effectiveness of models and methods that are the subject of theoretical research through numerical methods such as Monte Carlo simulations

Examples of models and methods subject to theoretical research

1. Statistical models and methods

  • Regression models

  • Generalized linear models

  • Time series models

  • Spatial models

  • Causal inference

  • Bayesian inference

2. Machine learning

  • Decision trees

  • Deep learning

  • Reinforcement learning

Eligible students

Type A students, Type B students, and PEARL students are all eligible to apply.

Application guidelines

Please check the program guide for details on application guidelines and contact information for the DEEP coordinator.

  • Application period: In principle, after the announcement of grades immediately following fulfillment of the DEEP completion requirements. However, if an application is not submitted at the end of the third year, it may be submitted at the end of the fourth year. In addition, as long as the DEEP completion requirements are met, applications will be accepted even if the student does not graduate by the end of the fourth year.

  • How to apply: Use the designated form on K-Support (details will be posted when grades are announced)

  • Documents to be submitted: Academic transcript, coursework

Contact information for inquiries

DEEP coordinator