Data-Driven Decisions and Digital Economics - 2nd year of master's degree

Data Analytics - 12 ECTS

• Machine Learning
  Close

  Machine Learning

  Ects : 6

  Lecturer :

  • FABRICE ROSSI

  Total hours : 36

  Overview :

  The course gives a thorough presentation of the machine learning field and follows this outline:

  1. general introduction to machine learning and to its focus on predictive performances (running example: k-nearest neighbours algorithm)
  2. machine learning as automated program building from examples (running example: decision trees)
  3. machine learning as optimization:
     1. empirical risk minimization
     2. links with maximum likelihood estimation
     3. surrogate losses and extended machine learning settings
     4. regularisation and kernel methods (support vector machines)
  4. reliable estimation of performances:
     1. over fitting
     2. split samples
     3. resampling (leave-one-out, cross-validation and bootstrap)
     4. ROC curve, AUC and other advanced measures
  5. combining models:
     1. ensemble techniques
     2. bagging and random forests
     3. boosting
  6. unsupervised learning:
     1. clustering (hierarchical clustering, k-means and variants, mixture models, density clustering)
     2. outlier and anomaly detection

  Coefficient : 2 6 (M2 Economie Internationale et Développement) 6 (M2 Diagnostic économique international)

  Require prerequisites :

  • intermediate level in either Python or R. Students are expected to be able to perform standard data management tasks in Python or R, including, but not limited to:
    • loading a data set from a CSV file
    • recoding and cleaning the data set
    • implementing a simple data exploration strategy based on pivot table and on graphical representation
  • intermediate level in statistics and probability. Students are expected to be familiar with:
    • descriptive statistics
    • conditional probabilities and conditional expectations
    • core results from statistics: bias and variance concepts, strong law of large numbers, central limit theorem, etc.

  Learning outcomes :

  After attending the course the students will

  • have a good understanding of the algorithmic and statistical foundations of the main machine learning techniques
  • be able to select machine learning techniques adapted to a particular task (exploratory analysis with clustering methods, predictive analysis, etc.)
  • be able to design a model selection procedure adapted to a particular task
  • report the results of a machine learning project with valid estimation of the performances of their model

  Assessment :

  • quizzes and tests during the course
  • machine learning project
• Machine Learning
  Close

  Machine Learning

  Ects : 6

  Lecturer :

  • FABRICE ROSSI

  Total hours : 36

  Overview :

  The course gives a thorough presentation of the machine learning field and follows this outline:

  1. general introduction to machine learning and to its focus on predictive performances (running example: k-nearest neighbours algorithm)
  2. machine learning as automated program building from examples (running example: decision trees)
  3. machine learning as optimization:
     1. empirical risk minimization
     2. links with maximum likelihood estimation
     3. surrogate losses and extended machine learning settings
     4. introduction to deep learning
  4. reliable estimation of performances:
     1. over fitting
     2. split samples
     3. resampling (leave-one-out, cross-validation and bootstrap)
     4. ROC curve, AUC and other advanced measures
  5. combining models:
     1. ensemble techniques
     2. bagging and random forests
     3. boosting
  6. machine learning for causal inference

  Coefficient : 1

  Require prerequisites :

  • intermediate level in either Python or R. Students are expected to be able to perform standard data management tasks in Python or R, including, but not limited to:
    • loading a data set from a CSV file
    • recoding and cleaning the data set
    • implementing a simple data exploration strategy based on pivot table and on graphical representation
  • intermediate level in statistics and probability. Students are expected to be familiar with:
    • descriptive statistics
    • conditional probabilities and conditional expectations
    • core results from statistics: bias and variance concepts, strong law of large numbers, central limit theorem, etc.

  Learning outcomes :

  After attending the course the students will

  • have a good understanding of the algorithmic and statistical foundations of the main machine learning techniques
  • be able to select machine learning techniques adapted to a particular task
  • be able to design a model selection procedure adapted to a particular task
  • report the results of a machine learning project with valid estimation of the performances of their model

  Assessment :

  • quizzes and tests during the course
  • machine learning project
• Data Science Project
  Close

  Data Science Project

  Ects : 3

  Total hours : 18

  Coefficient : 0,5
• Time Series and Anomaly Detection
  Close

  Time Series and Anomaly Detection

  Ects : 3

  Lecturer :

  • MADALINA OLTEANU

  Total hours : 24

  Overview :

  This lecture is thought as an introduction to the analysis of complex data, and particularly to that having a temporal component. Methods aimed at exploring and modelling time series, longitudinal data and graphs with temporal components will be addressed. The issues of detecting patterns, breakpoints, changes of regimes, and anomalies will be at the core of the different approaches. The first chapters will be devoted to hidden Markov models. After having briefly recalled some definitions and properties of Markov processes, we will define hidden Markov processes, illustrate them with several examples and give some of their properties. Inference techniques using the EM algorithm and Bayesian approaches will be presented and illustrated in practice. We will particularly focus on some specific models which are extremely useful for segmenting time series stemming from the economics field, such as autoregressive Markov switching models. The second part of the lecture will tackle the issue of change-point detection methods. We will start by introducing the change-point detection issue. More specifically, we will consider several frameworks and derive inference procedures for computing and locating change-points : online vs. offline strategies, single vs. multiple change point detection, known vs. unknown number of change points, parametric vs. non-parametric approaches. The third chapter will be aimed at introducing the issue of anomaly detection in the context of temporal data. After having defined what an anomaly is, we will start by assessing whether and how hidden-Markov models and change-point analysis may be useful for detecting anomalies. Then, we will compare these two approaches with other techniques, stemming either from the field of computational statistics, or from that of machine learning. During this chapter, we will also consider the questions of detecting patterns and clustering temporal data. The fourth chapter will address data that can be modelled as a graph or a temporal graph. We will start b y introduce some definitions and summaries for characterising the network (degree distribution, centrality indices, ...). Afterwards, we will tackle the questions of community detection and graph clustering. Eventually, we will address the issues of random networks and associated tests for randomness. Models and methods introduced in this lecture will be practiced using existing implementations in R and "real-life" datasets.

  Coefficient : 0,5

  Require prerequisites :

  Students are expected to have some notions of probabilities, statistical inference theory and time series analysis (ARIMA models). An intermediate

  knowledge of R and/or Python is also desirable.

  Learning outcomes :

  Gain some background and perspective of time series analysis from a data science point of view.

  Be able to handle temporal data subject to anomalies and change-points.

  Have some basic knowledge about graphs and temporal graphs mining.

  Assessment :

  Data challenge. A project to be done individually or by two, analysing real life data.

Digital Economics - 18 ECTS

• Competition and network economics
  Close

  Competition and network economics

  Ects : 3

  Lecturer :

  • ANNA CRETI
  • ANTOINE CHAPSAL

  Total hours : 30

  Overview :

  Theory and practice of competition in network industries; antitrust issues; theory of network and network effects; two-sided platforms and pricing

  Coefficient : 0,5

  Recommended prerequisites :

  Industrial Organization

  Require prerequisites :

  Advanced Micro

  Learning outcomes :

  Understanding of competition and regulation issues in network and digital economics

  Assessment :

  Written exam

  Bibliography-recommended reading

  Belleflamme-Peitz, Industrial Organization

• Blockchain economics
  Close

  Blockchain economics

  Ects : 6

  Lecturer :

  • LOUIS BERTUCCI

  Total hours : 36

  Overview :

  While this is a fairly recent technology, this class will take students through the fundamentals of blockchains as well as implications regarding financial, economic or social interactions. The class will start by some history needed to understand what lead to the creation of Bitcoin, the first blockchain, in 2009. We will then review the detailed functioning of a blockchain. We will continue by discussing important current developments in the industry as well as implications for the economic environment. Lastly, we will discuss potential future developments and how blockchains will impact a broad range of industries. Students will also be introduced to recent academic work related to blockchains.

  Students will be asked to pick a blockchain project from a list and present it briefly during the presentation session in front of the class (group presentation). When reaching this presentation session, students will be expected to be able to assess the pros and cons of a given blockchain project, and have a critical opinion on this project.

  Coefficient : 1

  Recommended prerequisites :

  The first prerequisite is coding. Knowledge in Python and/or Javascript will greatly help students perform the homework. Student less familiar with Python are expected to increase their Python skills by the end of the semester.

  The second prerequisite is basic economics (competition, market economy, utility maximization).

  While knowledge in computer science and economics is needed to properly understand what a blockchain is, we will go through what is needed just to make sure everyone is on the same page. In particular we will go through asymmetric cryptography, distributed networks, consensus, game theory, financial markets and corporate finance. Although students with knowledge in any of those topics will be more confortable, I intend to present them “from scratch”.

  Learning outcomes :

  The objective of this class is to give students a deep theoretical overview of what a blockchain is. Nonetheless we will also use mock-blockchains, write smart contracts and interact with them, through some computer sessions. This will help solidify the knowledge learned and de-mystify the functioning of a blockchain.

  Students will gain a deep knowledge of how a blockchain works internally. They will also be very aware of the different issues and perhaps they will be able to spot new use cases for a blockchain.

  Assessment :

  The evaluation is composed of a group presentation (1/3), homework (1/3) and a final exam (1/3). Class participation can be highly rewarded especially for students who struggle with homework. Students are encouraged to actively interact during the class.

  Bibliography-recommended reading

  Melanie Swan, Blockchain: Blueprint for a new economy, O’Reilly, 2015

  Andreas Antonopoulos, Mastering Bitcoin, 2nd edition, O’Reilly, 2017

  Andreas Antonopoulos / Gavin Wood, Mastering Ethereum, 1st edition, O’Reilly, 2018

  Primavera De Filippi/ Aaron Wright, Blockchain and the Law: The Rule of Code, Harvard University Press, 2018

• Financial Data et Systemic risk
  Close

  Financial Data et Systemic risk

  Ects : 3

  Lecturer :

  • Sylvain BENOIT

  Total hours : 24

  Overview :

  The course will equip students with the necessary knowledge to be able to undertake econometric analysis of the type commonly associated with modern financial econometrics research. Substantial emphasis will be placed on the development of programming skills in Python (or in MATLAB, especially for financial contagion and multivariate analysis).

   

  Course outline:

  1. Data collection (CRSP-Compustat, Yahoo-Finance, ECB data warehouse)
  2. Market Risk Measurement (Value-at-Risk, Expected Shortfall) – ARCH/GARCH models – univariate time series
  3. Backtesting tests for market-risk measurement (independence test, unconditional coverage test, conditional coverage test, super exception)
  4. Systemic Risk and Macroprudential regulation (SIFIs identification, MES, SRISK, ?CoVaR) – multivariate time series
  5. Principal Component Analysis (absorption ratio computation)
  6. Contagion models (direct and indirect effects decomposition)

  Coefficient : 0,5

  Recommended prerequisites :

  Time Series Analysis. Python programming.

  Learning outcomes :

  The course provides a deep knowledge of the advanced time series techniques and their application to systemic risk. A technical presentation of these models will be given, before studying applications of these models to systemic risk.

  Assessment :

  Individual homework assignment.

  Bibliography-recommended reading

  Benoit, S., Colliard, J.-E., Hurlin, C. and C. Pérignon (2017) Where the Risks Lie: A Survey on Systemic Risk, Review of Finance, 21(1), 109-152.

  Benoit, S., Hurlin, C. and C. Pérignon (2019) Pitfalls in Systemic-Risk Scoring, Journal of Financial Intermediation, 38, 19-44.

  Campbell, S. D. (2004) A Review of Backtesting and Backtesting Procedures, Working paper, Federal Reserve Board.

  Christofferson, P. and Pelletier, D. (2004) Backtesting Value-at-Risk: A Duration-Based Approach, Journal of Financial Econometrics, 2(1), 84-108.

  Du, Z. and J. C. Escanciano (2015) Backtesting Expected Shortfall: Accounting for Tail Risk, Management Science.

  Diebold, F.X. and K. Yılmaz (2009) Measuring Financial Asset Returns and Volatility Spillovers, with Application to Global Equity Markets. The Economic Journal, 119(1), 158-171.

  Diebold, F.X. and K. Yılmaz (2012) Better to Give than to Receive: Predictive Directional Measurement of Volatility Spillovers, International Journal of Forecasting, 28(1), 57-66.

• Experimental Economics
  Close

  Experimental Economics

  Ects : 3

  Lecturer :

  • CLAIRE RIMBAUD

  Total hours : 21

  Overview :

  The module will cover both methodology - why and how experiments in economics are conducted - and specific topics from the experimental literature via recent research articles.

  Coefficient : 2 pour le M2 296 et 0,5 pour le M2 346

  Learning outcomes :

  The aim of the module is to introduce students to the use of experimental methods in economics.

  Bibliography-recommended reading

  Charness, G., & Pingle, M. (Eds.). (2021). The art of experimental economics: twenty top papers reviewed. Routledge.

  Friedman, D., & Sunder, S. (1994). Experimental methods: A primer for economists. Cambridge university press.

  Moffatt, P., Starmer, C., Sugden, R., Bardsley, N., Cubitt, R., & Loomes, G. (2009). Experimental economics: Rethinking the rules. Princeton University Press.

  + articles cited in class.

• Private Cryptocurrencies
  Close

  Private Cryptocurrencies

  Ects : 3

  Lecturer :

  • ROMAIN PLASSARD

  Total hours : 21

  Overview :

  The first part of the course characterizes the long-standing project of creating an electronic cash system, analyzes the technical and institutional problems raised by its implementation, and reviews the proposed solutions leading up to Bitcoin. The second part examines the political motivations underlying the design of electronic cash systems.

  Coefficient : 0,5

  Learning outcomes :

  This course examines the origins and development of private cryptocurrencies. By the end of the course, students should be able to:

  • Identify the key challenges involved in designing electronic cash systems (e.g., double-spending prevention).
  • Describe the cryptographic mechanisms used to secure electronic payments (e.g., blind signature schemes).
  • Explain the conditions that enabled the emergence of Bitcoin and subsequent alternatives (e.g., Ripple, Ethereum, ZCash).
  • Assess and justify the design choices of different electronic cash systems.

  Assessment :

  Quizz (40%) and oral presentation (60%).

  Bibliography-recommended reading

  • Brunton, F. (2019). Digital Cash: The Unknown History of the Anarchists, Utopians, and Technologists Who Created Cryptocurrency. Princeton University Press.
  • Chaum, D. (1983). “Blind Signatures for Untraceable Payments.” Advances in Cryptology, Proc. Crypto ’82, D. Chaum, R. L. Rivest, & A. T. Sherman (eds.), Plenum Press, New York, pp. 199–203.
  • Chaum, D. (1985). “Security Without Identification: Transaction Systems to Make Big Brother Obsolete.” Communications of the ACM, 28, pp. 1030–1044.
  • Nakamoto, S. (2008). “Bitcoin: A Peer-to-Peer Electronic Cash System.” https://www.metzdowd.com/pipermail/cryptography/2008-October/014810.html
  • Turner, F. (2006). From Counterculture to Cyberculture: Stewart Brand, the Whole Earth Network, and the Rise of Digital Utopianism. University of Chicago Press.

Data Analytics - 12 ECTS

• NLP for economic decisions
  Close

  NLP for economic decisions

  Ects : 3

  Total hours : 24

  Overview :

  General introduction: aims and uses of NLP

   

  Part I: translating texts into quantities

  1. Tokenization, preprocessing
  2. Word frequency and inverse-document frequency (tf and tf-idf)
  3. Word embeddings

   

  Part II: Traditional NLP methods

   

  Methods used before LLMs

  1. Text classification
  2. Topic modelling and text summarization
  3. Sentiment analysis

   

  Part III: NLP with LLMs

  1. Short presentation of LLMs: transformers, encoders-only models, decoders-only models
  2. Text classification with LLM
  3. Topic modelling and text summarization
  4. Sentiment analysis
  5. Generating prompt for generative LLM
  6. Fine-tuning of LLM

  Coefficient : 2 pour le M2 296 et 0,5 pour le M2 346

  Recommended prerequisites :

  Basic knowledge of Python, knowledge of classifications methods in Data science

  Learning outcomes :

  At the end of this course, the students should:

  • have a good knowledge of the main methods used in Natural Language Processing
  • be able to implement these methods to data with Python
  • be able to interpret the results obtained through these examples

  Assessment :

  The evaluation is based on a project made by groups of two students

  Bibliography-recommended reading

  • Albrecht, Jens, Ramachandran, Sidharth, Winkler, Christian. Blueprints for text analytics using Python machine learning-based solutions for common real world (NLP) applications, O'Reilly, 2020
  • Alammar Jay and Maarten Grootendorst. Hands-On Large Language Models, Language Understanding and Generation, O'Reilly, 2024
  • Useful website: Kaggle
• Machine Learning for Economists
  Close

  Machine Learning for Economists

  Ects : 3

  Lecturer :

  • Mathilde GODARD

  Total hours : 24

  Overview :

  Economic science has evolved over several decades toward greater emphasis on empirical work. Ever increasing mass of available data (’big data’) in the past decade is likely to have a further and profound effect on economic research (Einav and Levin, 2014). Beyond economic research, governments and the industry are also increasingly seeking to use ’big data’ to solve a variety of problems, usually making use of the toolbox from machine learning (ML).

  The question we ask in this course is the following : What do we (not) learn from big data and ML as economists? Is ML merely applying standard techniques to novel and large datasets? If ML is a fundamentally new empirical tool, how does it fit with what we know? In particular, how does it fit with our tools for causal inference problems? As empirical economists, how can we use big data and ML? We’ll discuss in detail how ML is useful to collect new data, for prediction in policy, and to provide new tools for estimation and inference.

  Coefficient : 2 pour le M2 296 et 0,5 pour le M2 346

  Recommended prerequisites :

  Python (beginner/intermediate), Machine Learning, Microeconometrics.

  Learning outcomes :

  Course objectives:

  1. Present a way of thinking about ML that gives it its own place in the econometric toolbox.

  2 Develop an intuition of the problems to which it can be applied, and its limitations.

  3. Think of unstructured data (text, image) as data we can use when economic outcomes are missing.

  4. Specific focus on application of ML to social policies (health/labor/taxation/environment etc.).

  Assessment :

  Grading:

  1. In-class pairwise presentation of an academic paper (30% of overall grade).

  2. Final exam (in-class written text). 70% of overall grade.

  Bibliography-recommended reading

  • Mullainathan, Sendhil and Jann Spiess (2017). “Machine learning: An applied econometric approach”. In: Journal of Economic Perspective 31.2, pp. 87-106.
  • Kleinberg, Jon et al. (2015). “Prediction policy problems”. American Economic Review 105.5, pp. 491-495.
  • Athey, S. (2017): “Beyond prediction: Using big data for policy problems”, Science 355, 483–485.
  • Kleinberg, J., Lakkaraju, H., Leskovec, J., Ludwig, J. and S. Mullainathan (2018): “Human Decisions and Machine Predictions”, The Quarterly Journal of Economics, Volume 133, Issue 1, Pages 237–293.
  • Susan Athey, Guido W. Imbens. 2019. Machine Learning Methods That Economists Should Know About. Annual Review of Economics 11:1, 685-725.
  • Athey, Susan, and Guido Imbens. 2016. “Recursive Partitioning for Heterogeneous Causal Effects”. PNAS 113(27): 7353–60.
  • Belloni, A., V. Chernozhukov, S. Mullainathan and J. Spiess and C. Hansen.(2014): “High-Dimensional Methods and Inference on Structural and Treatment Effects” Journal of Economic Perspectives, Volume 28, Number 2 – Spring 2014, Pages 29–50
• Neural Networks
  Close

  Neural Networks

  Ects : 3

  Lecturer :

  • JOSEPH RYNKIEWICZ

  Total hours : 18

  Coefficient : 0,5
• Data visualisation
  Close

  Data visualisation

  Ects : 3

  Total hours : 15

  Coefficient : 0,5

Digital Economics - 9 ECTS

• Empirical Industrial Organization
  Close

  Empirical Industrial Organization

  Ects : 3

  Lecturer :

  • Daniel HERRERA ARAUJO

  Total hours : 21

  Coefficient : 2 pour le M2 296 et 0,5 pour le M2 346
• Solidity and smart contract development
  Close

  Solidity and smart contract development

  Ects : 3

  Lecturer :

  • TIANCHAN DONG

  Total hours : 18

  Overview :

  Blockchain is an amalgamation of several existing technologies. This course begins with an introduction to the historical developments, technologies, and ideologies that led to the emergence of blockchain. Students will then explore the layered architecture of blockchain systems and analyze the key technical and economic factors influencing each layer. To illustrate these concepts, smart contracts will be used to programmatically simulate system behaviors and design mechanisms. By the end of the first half of the course, students will have completed the full smart contract development lifecycle using the Remix IDE.

  The second half of the course adopts a case study approach to examine real-world blockchain applications. Participants will apply both analytical reasoning and the technical skills developed earlier in the course to formulate consultative assessments and solution strategies for contemporary blockchain use cases.

  Coefficient : 0,5

  Recommended prerequisites :

  Familiarity with Remix IDE.

  Basic understanding of computer concepts such as data structure and networking.

  Require prerequisites :

  Mastery in at least one programming language. Preferably Javascript.

  Fundamental theory of blockchain presented in the previous semester.

  Learning outcomes :

  By the end of this course, participants will demonstrate the ability to:

  1. Interpret and articulate blockchain system architecture, including peer-to-peer networking, consensus mechanisms, and application-layer components.
  2. Assess the technical robustness of blockchain projects.
  3. Assess the security and economic aspects of smart contracts.
  4. Implement core Ethereum standards and protocols in practical contexts involving cryptocurrencies, NFTs, and DAOs.
  5. Collaborate effectively in open-source environments by utilizing Github and Remix tooling, and community-driven development practices.

  Assessment :

  1. A weekly, 15min group presentation which expands on lecture topics.
  2. Bi-weekly homework projects.
  3. Final DAO project.

  Bibliography-recommended reading

  Solidity Documentation - docs.soliditylang.org/en/latest/ Code examples - solidity-by-example.org

• Platform economics
  Close

  Platform economics

  Ects : 3

  Total hours : 18

  Coefficient : 0,5

Job Market Insertion - 9 ECTS

• Business Cases
  Close

  Business Cases

  Lecturer :

  • RENE AID

  Total hours : 24

  Overview :

  The lecture is a sequence of use-cases in the industry and in the service business performed by professionals.

  Coefficient : Validation

  Learning outcomes :

  Knowing some of the most common use cases of data sciences in firms decision making business.

  Assessment :

  Presentation of a resume of the lectures before a jury.

• Communication
  Close

  Communication

  Total hours : 12

  Coefficient : Validation
• Internship
  Close

  Internship

  Ects : 9

  Overview :

  Students must intern at a company (semester 4) for at least 4 months.

  Coefficient : 1,5

  Learning outcomes :

  The students should also complete an end-of-studies internship lasting at least 4 months. The curriculum includes guest lectures by visiting professionals on issues related to Big Data, providing another means of connecting with relevant business circles.

  Assessment :

  The internship will conclude with a report to be reviewed by a committee.