We have already reached the fifth anniversary of the publication of the Courrier des Statistiques journal in its new format, with over 70 articles published, still with a focus on education and the desire to open up to new subjects, new services etc., with the number of issues now in double digits and a modernised model.

This issue, N10, begins with a subject that has never been addressed in the new Courrier (but featured multiple times in the previous one), that of the dissemination of statistics. More precisely, the subject in question is datavisualisation (or “dataviz” for those in the know). The article by Christine Lagarenne, Frédéric Minodier and Odile Samson is rich and provides numerous examples, including ancient ones (the extraordinary map by Charles Joseph Minard on the Napoleonic invasion of Russia), allowing us to touch on every aspect of this activity. However, it should not be hastily summarised as the creation of “pretty” charts. Sitting at the crossroads between dissemination and communication, datavisualisation not only seeks to simplify messages to allow an understanding at a glance, but it must also make readers want to read. To achieve this, it uses different techniques: infographics, narrative scrolling, interactive datavisualisation etc. The article also raises operational, technical and organisational questions, on a topic that is now unavoidable for statisticians.

One of the Courrier’s constant preoccupations is ensuring that it gives all the Ministerial Statistical Offices a platform, in order to show how statistics are applied in different sectors of activity. So far, the Courrier has published articles from the MSOs of the Ministry of Justice (issue N1), the Ministry of the Interior (issue N2 and N7), the Ministry of Health (issue N4), the Service des données et études statistiques (Data and Statistical Studies Service, SDES) for housing (issue N4), the Ministry of Education (issue N5, N6 and N7), the Ministry of Agriculture (issue N7) and the Ministry of Local Authorities (issue N8). In this issue, two new subject areas are discussed, defence and sport.

In the article on defence statistics, Pierre Greffet discusses the specific characteristics of the field, which is not subject to European regulations, is not easily based on the general nomenclatures (NAF) and maintains regular and highly formalised links with the world of research. He also highlights a paradox: such data are often highly confidential, with a level of sensitivity that goes beyond statistical confidentiality, while at the same time there is a desire for open access, especially for researchers. We can resolve this apparent contradiction by stressing that while some subjects may be sensitive (defence economics), others are not (visits to places of remembrance); thus, a project aimed at organising a highly controlled opening up of access is under consideration, incorporating the principle of a data room.

As for statistics on sport, this is a subject that has no framework imposed by international regulations and is, in fact, poorly framed in general. Augustin Vicard presents the sources available in this area and highlights the limitations of administrative data and the challenges of describing the concept of participation in sports, as well as the difficulties related to the plethora of sports, some of which are more obscure. Conducting surveys on the subject therefore requires firm decisions, for example by choosing to focus on regular participation in sports. Beyond the production process, the very question of the challenges entailed in these statistics arises: do we consider sport to be a social phenomenon? Do we consider it to be a physical activity, with a particular health aspect, and public policies that go with it? Finally, the author addresses the question of data from connected applications and sensors, which are interesting but which do not allow for standardised monitoring of participation.

This is followed by a series of two articles with very similar structures, this time on the fields of education and health, but which are not strictly speaking articles by the MSOs: in fact, they do not concern statistics but registers. From this point of view, they echo the dossier on registers in issue N8, which contained articles on the Répertoire national d’identification des personnes physiques (National Register for the Identification of Natural Persons, RNIPP), the Système national de gestion des identifiants (National Identification Management System, SNGI), the Répertoire d’unités statistiques (Statistical Business Register, SIRUS), and the Base permanent des équipements (Permanent Equipment Database). In this issue, two registers of establishments are described: the Répertoire des établissements sanitaires et sociaux (Register of Health and Social Establishments, FINESS) and the Répertoire des établissements du système éducatif (Academic and Ministerial Register of Establishments in the Education System, Ramsese). It is highly recommended that readers read the two articles “in tandem” to identify the aspects in common and the differences.

FINESS is a well-known register that was created more than 40 years ago (1979) and which plays a fundamental role in the regulation, assessment, management, financing and identification of the structures that it includes. It is characterised by a very demanding formal framework, with the registration of any data requiring the existence of legal or administrative acts. Each establishment has an identification number. FINESS plays a major role in the health information systems ecosystem. It is linked with two other registers in the field of health, the Répertoire partagé des professionnels de santé (Shared Register of Health Professionals, RPPS) and the Répertoire opérationnel des ressources (Operational Resources Register, ROR), and is paired with the Sirene business register. It is widely used by central government bodies, the general public and the establishments themselves. Its central role and its broad dissemination mean that it is subject to stringent quality requirements. In order to alleviate certain limitations, it is currently being redesigned in a project led by the Agence du numérique en santé (Agency for Digital Technology in Health).

Created in 1977, around the same time as FINESS, the Répertoire académique et ministériel sur les établissements du système éducatif (Academic and Ministerial Register of Establishments in the Education System, Ramsese) is based on a very different approach, taking advantage of the regional organisation of the education system. Thus, the register is handled in a decentralised manner by the Academic Statistical Offices (ASOs), with the work of clerks registers at local level being decisive in guaranteeing the quality of the data in the register. It has a wide range of uses which allow users to meet statistical, governance, management or interoperability needs, with each establishment having a unique identification number. Ramsese plays a central role in the education system and ensures the consistency of data from public structures using the Sirene business register. Its API-based dissemination facilitates the sharing of data in the context of the urbanisation projects of the applications. Ramsese’s visibility is expanding, with some of its data being made available in open data format.

This issue ends with a subject area that is, in principle, more traditional in the world of statistics, that of survey sampling. Well, it is traditional… to a certain point, because Pascal Ardilly addresses a subject that is little discussed within official statistics: non-probability sampling. While non-probability sampling is not, in principle, based on the random selection of a sample, it is the norm within survey institutes. Using an educational approach, the author sets out the differences between random and non-probability sampling, with a number of diagrams. Non-probability sampling, which is effective in terms of controlling costs, is less effective in controlling errors, with a specific problem relating to bias in particular. However, there is also a genuine theoretical justification for using it. Finally, using two examples, the article sheds new light on the use of big data, delivering a message: warning, quantity is not a guarantee of quality…