Abstract: Graphs provide interesting models for representing interactions between entities.
While a large part of the literature is dedicate to the simple situation where interactions are static and binary,
many real world situations ask for more complex models which account for the temporal dimension of the interactions
as well as for their other characteristics. In this joint work with M. Corneli, P. Latouche and C. Bouveyron,
we propose extensions of the stochastic block model to interaction data with a temporal component and with a textual one.
A classical way of representing temporal interactions consist in constructing a time series of graphs, where each graph aggregates the interactions over a given time interval. While this approach can produce interesting results, it cannot adapt to more complex schemes where a single time scale cannot capture the full temporal dynamic of the interactions. In our approach, we use a continuous time model which enables us to limit temporal aggregation to very short time intervals, mostly motivated by computational consideration. We propose in addition a variant of our model that can model text messages attached to interactions via a dynamic topic model
The presentation will begin with an introduction to generative models for graphs with a gradual introduction of elements needed to handle continuous time and information attached to interactions. The models will be illustrated on two real world examples: a validation study using the bike sharing system of London and a more complex analysis of the Enron email database.
Bio of the Speaker: Fabrice Rossi is full Professor of Data Science at University Paris Dauphine, PSL. He is a member of the CEREMADE and of the MILES team. He specialises on exploratory data analysis with a special interest in graph data, change detection and visual data exploration. More generally, his research covers numerous important themes of machine learning including large scale data processing, feature selection, learning theory and clustering. Fabrice Rossi works frequently with researchers from other fields, especially from the humanities, including archaeology, history and sociology. He has (co)-authored more than 160 articles in journals and conference proceedings.
If you’d like us to provide a sandwich for you, please confirm your attendance before Friday 22th November 2019
Abstract:This research is about the identification of management paradigms. It attempts to investigate the evolution of focus and themes in management by examining articles from the leading French newspapers over the last 20 years with cutting-edge methods in machine learning. Among available approaches, the probabilistic topic modelling is, perhaps, the most likely able to discover the thematic structure of texts. This method allows us to characterize topics in a large corpus of unannotated and unstructured documents. We completed the analysis using a Word2vec language processing approach. By doing so, we highlighted emerging (and disappearing) management topics throughout the years and how these topics have been characterized in terms of the kind of lexical dictionary that has been used in examined papers. (Co-authored with Svitlana Galeshchuk and Adam-Ledunois Sonia)
Short Bio of the Speaker: Sébastien Damart is full professor in Dauphine Center for Management Research (DRM) in University Paris Dauphine, PSL. His research has largely focused on topics related to instrumentation used to facilitate group work in organizations. He participated in the development of several research questions, within this thematic framework, whose final ambition was to devise approaches and tools that would allow a better integration of the logics and points of view of the members of a collective (a group, an organization, any territory). He is associated to Observatoire de l'innovation managériale. He is also an associate editor for Journal of Management History.
Abstract: Turn your Hobby into a Business” is one of the taglines displayed on the web platform Etsy. Created in 2005, this platform is dedicated to the sale of handmade products (clothes, decorative objects, jewels…). Today, it gathers 1.8 million sellers and 29.7 million buyers. Claiming its affiliation to the “gig economy” and advocating a “social safety net that works for everyone who works”, Etsy encourages everyone who practices handicrafts to turn themselves into professional creative entrepreneurs by selling their items on the platform. To do so, they should focus on craft while the platform would help them to market their production.
Actually, the platform’s promise of a new source of (self-)employment via the professionalization of a hobby has not become a reality for most of Etsy “creators”. Only a minority earns a living thanks to the platform. Besides, this minority is not made of craftsmen or craftswomen but rather of designers who outsource the production. Moreover, the actual key to succeed on the platform is not to be a good craftsman but to invest time and effort on “digital labor” and finally to act as a mere “entrepreneur”.
Based on an empirical research, the aim of the presentation is to lay the emphasis on the various uses and the diverse kinds of work generated by a lucrative platform such as Etsy. In a sociological perspective, this issue is discussed though the analysis of social class and gender (most of Etsy creators being women from the middle and the upper classes). Data is both qualitative and quantitative. By establishing a typology of Etsy creators, the presentation will question the blurring boundaries between work and play, and the increasing social value of entrepreneurship in modern capitalist societies.
Short Bio of the Speaker: Anne Jourdain is an associate professor in sociology at the Paris-Dauphine University and researcher at IRISSO (Interdisciplinary Research Institute in Social Sciences). Her research fields are economic sociology, sociology of work and sociology of art, with specific focuses on arts and crafts and on the digital platform Etsy.
Different types of institutions are involved in social finance. This paper studies the taxonomy and tradeoffs of those institutions. Our contribution is twofold. First, we map out and classify the full continuum of conceivable social finance institutions (SFIs), ranging from foundations offering pure grants to social banks supplying soft loans. The in-between category includes under-researched “quasi-foundations” granting loans requiring partial repayment. Second, we develop a model under which SFIs face information asymmetries and trade off social screening against social contributions, under a budget constraint depending on their funders’ generosity. We characterize SFIs’ optimal strategies and discuss whether they involve social screening of loan/grant applicants. The model establishes that quasi-foundations can be efficient vehicles for social finance, especially when social screening costs are relatively low. It also shows that social screening is indispensable in equilibrium when initiators of socially oriented projects can self-select themselves as applicants to SFIs.
(Authors: Simon Cornée, Marc Jegers, Ariane Szafarz)
Short Bio of the Speaker:
Marc JEGERS is a Professor of Management from Vrije Universiteit Brussel (VUB). His main research field is in Managerial Economics, the Management of Non Profit Organisations, Accounting, Hospital Accounting and Financial Management of Hospitals.
Abstract: Social networks are expected to matter for invention in cities, but empirical evidence is still puzzling. In this paper, we provide new results on urban patenting covering more than twenty years of European patents invented by nearly one hundred thousand inventors located in France. Elaborating on the recent economic literatures on peer effects and on games in social networks, we assume that the productivity of an inventor’s efforts is positively affected by the efforts of his or her partners and negatively by the number of these partners’ connections. In this framework, inventors’ equilibrium outcomes are proportional to the square of their network centrality, which encompasses, as special cases, several well-known forms of centrality (Degree, Katz-Bonacich, Page-Rank). Our empirical results show that urban inventors benefit from their collaboration network. Their production increases when they collaborate with more central agents and when they have more collaborations. Our estimations suggest that inventors’ productivity grows sublinearly with the efforts of direct partners, and that they incur no negative externality from them having many partners. Overall, we estimate that a one standard deviation increase in local inventors’ centrality raises future urban patenting by 13%. We also find that geographically close relations are up to two third more beneficial to inventors than distant ones.
Short Bio of the Speaker:
Laurent Bergé is a postdoc in economics at CREA (university of Luxembourg) since 2016. He joined CREA after a postdoc in statistics at University Paris Descartes and holds his PhD in economics from the university of Bordeaux (2015). His research looks at the the determinants of innovation, the mechanisms of network formation and the link between network position and performance. Beyond his research in economics, he also develops methods and algorithms (R packages).
Speaker's website: https://sites.google.com/site/laurentrberge/
Abstract: Existing work provides contrasted arguments and evidence regarding the effects of a corporate scandal on the rivals of an implicated firm: while competitors may suffer from a negative contamination effect, they may also benefit from a positive substitution effect. Building and extending prior works, we develop a model to predict how contamination and substitution interact and jointly affect firm value during the course of a corporate scandal. We test and find support for our arguments using unique data, including 1.2m tweets, on the Dieselgate scandal that affected the global automobile industry in 2015.
Short Bio of the Speaker:
Julien Jourdan is a professor of strategy at Université Paris-Dauphine (PSL), affiliated with the Dauphine Center for Management Re- search and the French National Center for Scientific Research. He received his PhD in Strategic Management from HEC Paris. His research focuses on the social evaluation of organizations and its strategic implications.
Speaker's website: https://sites.google.com/site/julienjourdan/
Abstract: As algorithm-based judgement is often expected to outperform human judgement, algorithms are widely used to assist human with decision-making and also influence human’s decision-making. However, extant laboratory experiments find mixed evidences that humans may or may not appreciate algorithm-based judgement. While field evidence is scant, we leverage a unique dataset provided by a leading online peer-to-peer lending marketplace in China, where historical algorithm-based investments are visibly distinguishable from human investments for each loan listing upon individual lenders’ manual investment. We therefore investigate how algorithm-based investment influences manual investment. We find that algorithm-based investments positively affects subsequent lenders’ manual investment amount. Specifically, it is the average amount of the observed algorithm-based investments that positively affects the manual investment amount but not the number of the algorithm-based investments. We show that the impact of the observed algorithm-based investment decreases with the lenders’ experiences of manual investment, which is in line with the notion of “algorithm appreciation”. We further explore the underlying mechanism by ruling out several alternative explanations.
Short Bio of the Speaker:
Xitong Li is an Assistant Professor in the department of Information Systems and Operations Management, HEC Paris, France. His recent research interests include the economic and social impacts of using online data/information, and innovative technologies using online data and services. His research has appeared, or is forthcoming, in Information Systems Research, MIS Quarterly, Journal of Management Information Systems, ACM Transactions on Internet Technology, ACM Transactions on Multimedia Computing Communications and Applications, IEEE Transactions on Engineering Management, IEEE Transactions on Automation Science and Engineering, IEEE Transactions on Systems, Man & Cybernetics, IEEE Communications Magazine, and other leading international journals and conference proceedings. His research has recently been granted by ANR AAPG France (solo PI, JCJC) for three years of 2018-2020. He won the OCIS Best Paper Award Finalist at the 77th Academy of Management Annual Meeting in 2017 and the Best Paper Award at the 46th Hawaii International Conference on System Sciences (HICSS) in 2013. He received his Ph.D. in management from the MIT Sloan School of Management, and his Ph.D. in engineering from Tsinghua University.
Speaker's website: http://www.hec.fr/Faculte-Recherche/Membres-de-la-faculte/LI-Xitong
Abstract: The Italian judicial system is notoriously slow, with an estimated backlog of 5 million cases. We use a sample of 903,660 court cases in Turin to study the role that various adjudication procedures play in judicial delay. We exploit plausibly exogenous variation in the procedures governing how judges rule on small claims and implement a quasi-experimental approach to estimate the causal effect of less restrictive procedures on judicial delay. For any claim valued below e1,100, judges do not need to provide formal legal justification for their decisions. Judges can rule based on “equità”, i.e., fairness, intuition or commonsense grounds. For cases valued above this threshold, judges do not have such flexibility. Our regression discontinuity estimates, which ex- ploit the variation in these adjudication procedures just above and just below this threshold, reveal that when judges are able to rule without providing legal justification, decisions are made nearly six months faster. We discuss the policy implications in the realm of small claims including methods to ease congestion in Italian courts and efforts to improve judicial performance more broadly.
Short Bio of the Speaker:
Alessandro Melcarne (PhD. In Law and Economics) has been working as Maître de Conférence at EconomiX, CNRS & Université Paris Nanterre since 2016. His research mainly focuses on the independence and efficiency of the European judiciary system. In particular, his cross-disciplinary studies between Law and Economics explore how the judiciary efficiency in Europe influence firm dynamics.
Alessandro is awarded of the 2017-2020 Young Researcher Grant by the French National Research Agency (ANR). His papers have been published on the European Journal of Law and Economics, Review of Law and Economics, Economic Notes, and etc.
Speaker's website: economix.fr/fr/page/melcarne-alessandro?tab=presentation
Abstract: The use of expert committees is common in many settings. A key concern is the potential for conflict of interest, particularly for members of committees that oversee regulated firms. However, ties to industry may be correlated with relevant expertise. We examine the voting behavior of members of the Food & Drug Administration's Advisory Committees (ACs), which make recommendations on new drug applications and other regulatory questions. Our work exploits a novel dataset that includes detailed information on each AC member, including their academic degrees, age, areas of expertise, and scientific contributions. We construct a measure of financial ties to industry using information disclosed in scientific publications authored by AC members, as well as those reported directly to the FDA and by the industry under the Sunshine Act. Advisors with industry ties generally have higher observable quality: they publish more, receive more NIH grants, etc. We estimate a structural model of voting that allows us to recover each member's skill and bias associated with financial ties to a drug's sponsor or its competitors. Our preliminary results indicate that financial ties to industry are associated with both an increased probability of voting in favor of a drug as well as higher skill.
Short Bio of the Speaker:
Margaret Kyle (PhD, MIT Economics) studies innovation, productivity and competition. She has a number of papers examining R&D productivity in the pharmaceutical industry, specifically the role of geographic and academic spillovers; the firm-specific and policy determinants of the diffusion of new products; generic competition; and the use of markets for technology. Recent work examines the effect of trade and IP policies on the level, location and direction of R&D investment and competition. She also works on issues of innovation and access to therapies in developing countries. Her papers have been published in various journals of economics, strategy, and health policy, including the RAND Journal of Economics, Review of Economics and Statistics, Journal of Law and Economics, Strategic Management Journal, Health Services Research, and Health Affairs.
Margaret currently holds the Chair in Markets for Technology and Intellectual Property at MINES ParisTech. She is an associate editor at the International Journal of Industrial Organization and a Research Fellow at the Centre for Economic Policy. She previously held positions at Carnegie Mellon University, Duke University, London Business School and the Toulouse School of Economics. She has also been a visiting scholar at the University of Hong Kong and Northwestern University.
Speaker's website: www.margaretkyle.net
Abstract: This paper proposes a three-step estimation strategy for dynamic conditional correlation models. In the first step, conditional variances for individual and aggregate series are estimated by means of QML equation by equation. In the second step, conditional covariances are estimated by means of the polarization identity and consistent estimates of the conditional correlations are obtained by their usual normalization. In the third step, the two-step conditional covariance and correlation matrices are regularized by means of a new non-linear shrinkage procedure and used as starting value for the maximization of the joint likelihood of the model. This yields the final, third step smoothed estimate of the conditional covariance and correlation matrices. Due to its scant computational burden, the proposed strategy allows to estimate high dimensional conditional covariance and correlation matrices. An application to global minimum variance portfolio is also provided, confirming that SP-DCC is a simple and viable alternative to existing DCC models for any application of interest.
Short Bio of the Speaker: Claudio Morana is Professor of Economics at the University of Milan-Bicocca (Italy). Before moving to Milan University, he was Assistant and then Associate Professor of Economics at the University of Eastern Piedmont (Novara), Lecturer at Heriot-Watt University (Edinburgh, UK) and Aberdeen University (Aberdeen, UK). He also was a Fulbright Research Scholar at Michigan State University (US) and consultant for the European Central Bank. He is a Fellow of the Center for Research on Pensions and Welfare Policies (CeRP, Collegio Carlo Alberto,Torino), Senior Fellow of the Rimini Center for Economic Analysis (RCEA, Rimini, Italy) and member of the Editorial Board of Econometrics, Heliyon, the Open Journal of Statistics and the Asia-Pacific Journal of Mathematics. His research interests include macro, financial and climate change econometrics, the macro-finance interface, the interconnection of financial and economic cycles. He has published over 70 articles in academic journals in the above topics, including the Journal of Econometrics, Journal of Empirical Finance, Journal of Banking and Finance, Journal of International Money and Finance, Journal of Economic Dynamics and Control, Economic Modeling, Computational Statistics and Data Analysis,International Journal of Forecasting and Energy Economics. He was born in Turin, graduated from the University of Turin, and received his Economics MSc from the University of Glasgow (Glasgow, UK) and PhD from the University of Aberdeen (Aberdeen, UK).
Abstract: Network analysis is devoted to the study of a set of relationships with the aim of understanding the structure generated by these relationships. In this presentation we shall present an application of graph theory and network analysis in the field of international relations, particularly the case of ratifications of multilateral treaties related to the environment (work done in collaboration with Pierre Mazzega, University of Toulouse and Ana Flavia Barros-Platiau, University of Brasília). Indeed we can first analyze the network of "immediate ratifications" where a treaty X will be in relation with another treaty Y if there is a country that has ratified Y just after X. We can also have a dual vision where the ratifications of multilateral environmental agreements induce a network of countries linked in pairs when one ratifies the same agreement right after the other. Further modeling may be presented either on other networks that can be analyzed or on the use of hypergraphs.
Short Bio of the Speaker: Romain Boulet is maître de conférences in mathematics at IAE Lyon school of management, University Lyon 3. His research interests are graph theory (especially algebraic graph theory) and network analysis with a particular interest in interdisciplinary collaborations. He has worked on networks of legal texts, networks of international relations between countries but also networks from history or geography.
Speaker's website: www.romainboulet.fr
Abstract: The fraction of women in economics has grown significantly over the last forty years. Yet, differences in research output between men and women remain large and persistent. These output differences are reflected in large network differences across the genders. Women have fewer collaborators, collaborate more often with the same coauthors, and a higher fraction of their coauthors are also coauthors of each other. Moreover, women coauthor a large share of their work and do so with more senior coauthors. Standard models of homophily and discrimination cannot account for these differences. We discuss how differences in risk aversion between men and women can explain them.
Short Bio of the Speaker: Anja Prummer is a Lecturer at Queen Mary University of London. Previously, she was a Postdoctoral Fellow at the Cambridge-INET Institute, University of Cambridge. She received her Ph.D. from the European University Institute, Florence. Her research interests are Social Networks and Political Economy.
Speaker's website: https://sites.google.com/site/anjaprummer/
Abstract: Differences in regulated pharmaceutical prices within the European Economic Area can be exploited by pharmacy retailers using parallel imports. Such provision decisions affect the sharing of profits in markets for prescription drugs, including the profitability of innovating pharmaceutical companies before patent expiry, when parallel trade is the unique source of upstream competition. We develop a structural model of demand and supply where retailers can choose the set of goods to sell to consumers, thus foreclosing the access to some drugs, in response to differences in profitability across products. On the supply side, retailers bargain over wholesale prices with the manufacturer and parallel traders. With detailed transaction data, we identify a demand model with unobserved choice sets using supply side conditions for optimal assortment decisions of pharmacies. Estimating our model, we find that retailer incentives play a significant role in fostering parallel trade penetration. Our counterfactual simulations show that the parallel imports of drugs have large implications for the distribution of industry profits. In particular, retailers gain at the expense of pharmaceutical companies, while parallel traders also gain but more modest profits. Finally, a policy preventing pharmacies in foreclosing the manufacturer’s product is shown to partially shift profits from pharmacists to both parallel traders and the manufacturer, and a reduction in the regulated retail price favors the manufacturer even more.
Short Bio of the Speaker: Professor Dubois is a Scientific Director of the Toulouse School of Economics and the Managing Editor of the International Journal of Industrial Organization.
Speaker's website: https://pierredubois.github.io/
Governance Analytics vous invite à une session de formation sur le logiciel Tableau (https://www.tableau.com/fr-fr)
La présentation sera en français et les places sont limitées.
Merci de vous inscrire sur le site web suivant si vous êtes intéressé(e) :
Les participants sont encouragés à installer le logiciel avant de venir à la formation afin de réaliser quelques exercices.
Many of the projects that responded to the call of research cooperation of Governance Analytics are related to the textual and/or Web data. However, project owners may not have a clear idea of what data are available or how they can get and process it. In this presentation, I will propose tools and methods that may help researchers identify potential research questions and methods based on the examples of ongoing Governance Analytics projects.
Bruno Chaves Coordinator of Governance Analytics
Alors que la fonction publique emploie très largement des femmes, il y a relativement peu d’analyses sur les inégalités de rémunérations entre les hommes et les femmes dans ce secteur. Ce point aveugle est en partie lié à l’idée que les règles institutionnelles de recrutement et de fixation des rémunérations dans la fonction publique ajoutées au rôle protecteur des syndicats laissent moins de place à des inégalités selon le sexe. Toutefois un "plafond de verre" est également observé dans la fonction publique en France. Dans cette présentation, nous appliquerons une méthode proposée par Gobillon et al. (2015) pour estimer cette différence d'accès aux rangs les plus élevés. Dans un deuxième temps, nous montrerons que de ce point de vue, la fonction publique est très proche des pratiques du secteur privé.
Titre original de l’article : « Egalité professionnelle entre les hommes et les femmes : des plafonds de verre dans la fonction publique ? »
Publié dans : Economie et Statistique, 2016, volume 488-489, p. 97-121
Téléchargement : http://laurent.gobillon.free.fr/page_web/articles/fremigacci_al_2016_genre.pdf
Dominique Meurs est Professeure à l'Université de Paris Ouest Nanterre La Défense et Chercheuse à EconomiX (CNRS UMR 7235) et chercheuse associée Ined
Introduction of Prof. Winkels: Radboud G.F. Winkels is associate professor in Computer Science and Law at the "Leibniz Center for Law" (LCL) of the Faculty of Law of the University of Amsterdam, the Netherlands and dean of the PPLE college.
He started his academic career as a student in psychology at the University of Amsterdam. In April 1987 he received his Master's Degree (with honours) in Artificial Intelligence. Since then he has been working as a full-time AI-researcher at the University of Amsterdam, first at the department of "Social Science Informatics" (now Human-Computer Studies Lab) and since March 1989 at the department of Law and Computer Science (now LCL), where he is now senior lecturer/associate professor. In this last capacity he is also a participant in the research school SIKS. His research deals with Intelligent Learning Environments, and Artificial Intelligence and Law.
Legal network analysis : Sources of Law form a large and growing network. Legislation is full of internal and external references, case law cites other case law and refers to legislation, legal doctrine and commentaries cite both case law and legislation and possibly other commentaries. This reserarch is aimed at trying to exploit this network structure to help (legal) practitioners and legal scholars. We address questions like: Can we use network features to determine authority of cases? Can we use it to assess the impact of planned legislative changes? Can we exploit the network structure to suggests useful sources of law to a practioner given a particular item, in focus? Can legal scholars use netweork analysis techniques to discover breaches in lines of thinking of courts?
Professor Winkels will give us a lecture on computational legal network analysis and disucss some applications in social sciences. Anyone interested in inter-disciplianary research on social sciences, legal studies and computer science should not miss it.
Speaker's website: http://www.leibnizcenter.org/~winkels/
In this paper, we aim to bring the debate on the global productivity slowdown – which has largely been conducted from a macroeconomic perspective – to a more micro-level. We show that a particularly striking feature of the productivity slowdown is not so much a lower productivity growth at the global frontier, but rather rising labour productivity at the global frontier coupled with an increasing labour productivity divergence between the global frontier and laggard (non-frontier) firms. This productivity divergence remains after controlling for differences in capital deepening and mark-up behaviour, suggesting that divergence in measured multi-factor productivity (MFP) may in fact reflect technological divergence in a broad sense. This divergence could plausibly reflect the potential for structural changes in the global economy – namely digitalisation, globalisation and the rising importance of tacit knowledge – to fuel rapid productivity gains at the global frontier. Yet, aggregate MFP performance was significantly weaker in industries where MFP divergence was more pronounced, suggesting that the divergence observed is not solely driven by frontier firms pushing the boundary outward. We contend that increasing MFP divergence – and the global productivity slowdown more generally – could reflect a slowdown in the diffusion process. This could be a reflection of increasing costs for laggard firms of moving from an economy based on production to one based on ideas. But it could also be symptomatic of rising entry barriers and a decline in the contestability of markets. We find the rise in MFP divergence to be much more extreme in sectors where pro-competitive product market reforms were least extensive, suggesting that policy weaknesses may be stifling diffusion in OECD economies.
We quantify how access to frontier knowledge affects the creation of ideas. We show that citing frontier knowledge is correlated with producing high-quality papers. Because this correlation may be driven by unobserved factors, we identify the causal effect of frontier knowledge by analyzing a sudden collapse of international scientific cooperation. We show that World War I and the subsequent boycott against Central scientists severely reduced the dissemination of international knowledge, including knowledge at the scientific frontier. We then estimate how the reduction of international knowledge flows affected the productivity of scientists. Specifically, we compare productivity changes for scientists who relied on frontier knowledge from abroad, to changes for scientists who relied on frontier knowledge from home. After 1914, scientists who relied on frontier knowledge from abroad published fewer papers in top science journals and produced less Nobel Prize-nominated research. Our results indicates that access to the very best research, the top 1%, is essential for scientific progress.
Speaker's website: https://sites.google.com/site/aleiaria82/home
Link to the paper: https://www.dropbox.com/s/d25bzozaptd3t6q/Knowledge%20Accumulation3_small.pdf?dl=0
This project is designed to develop a comprehensive database on the French stock markets since 1796, and to be extended to other kinds of data and to other European countries. In combination with other research infrastructures such as the SCOB database of the Antwerp University, it would meet the need for a benchmark infrastructure, first in France and then in Europe, to develop high-quality research by both scholars and the financial services.
Big data, smart data, data science... sont autant de termes qui ont diffusé auprès du public et des différents domaines scientfiques en quelques années. Les domaines de la santé et la recherche biomédicale en particulier s'y sont intéressé à partir de 2008, et le nombre de publications portant sur le "big data" en santé augmente exponentiellement depuis 2012. Néanmoins, nous sommes encore essentiellement au temps des promesses qui restent à concrétiser et à évaluer. En France, en particulier, il existe encore assez peu d'initiatives concrètes relevant d'une data science en santé. Parmi ces initiatives, la plupart sont actuellement portées par le milieu hospitalier, que ce soit en Ile de France (APHP) ou dans le Grand Ouest (CHRU de Brest, Rennes). Nous présentons ici le contexte du big data en santé, ainsi que deux projets actuellement en cours de déploiement en milieu hospitalier. Le projet d'entrepôt de données de santé soutenu par l'APHP et recouvrant les données de tous les établissements de l'APHP sera également présenté. D'autres projets en perspective à court et moyen termes seront esquissés.
Les intervenants :