Centres of Excellence on Time Series Analysis and Statistical Disclosure Control COSA
Project coordinator and Co-Partners
Coordinator : France (INSEE)
Co-Partners : Belgique (Banque nationale de Belgique (NBB)), Italie (Istat), Allemagne (Destatis),
Hongrie (HCSO), Espagne (INE), Pays-Bas (CBS), Islande (Statice), Slovénie (SURS),
Autriche (STAT), Norvège (SSB)
Start date : 01/11/2025 – End date : 31/10/2029
Objective
The COSA (Confidentiality, Seasonal Adjustment, and Time Series) project is structured around two Centers of Excellence (CoE): one dedicated to time series analysis (TSA), coordinated by INSEE, and the other dedicated to data confidentiality (SDC), coordinated by Statistics Netherlands (CBS). The Deutsche Bundesbank strengthens the time series component as a special partner. Other institutions will be involved as members of user groups and observers (ECB, ABS, OECD, etc.). Against the backdrop of developing common tools and methods within the ESS, this initiative will continue the work of the STACE project (2020–2024) along three main lines: knowledge sharing and support for practitioners, the development and maintenance of custom software (JDemetra+ for TSA, τ-ARGUS and sdcMicro for SDC), and methodological guidance (manuals, guidelines, research papers). The overall objective is to provide user-friendly, effective, and technically up-to-date tools for statistical production in these two areas. The planned work is divided into three work packages (WP), each with its own specific objectives, deadlines, and deliverables.
Distribution of work
- WP1 (coordinated by INSEE): The objective is project management, coordination and communication
- WP2 (coordinated by CBS): The objective is to strengthen support for practitioners and methodological advice, on the one hand, and to modernise the tools, in particular τ-ARGUS, starting with the development/design of a modern software architecture and implementing some of these ideas as part of a proof of concept. These developments will require INSEE to carry out maintenance work on the R package rtauargus, which in its current version launches τ-Argus on every call. The other SDC tools, as well as the ‘old’ τ-ARGUS, must continue to be maintained.
- WP3 (coordinated by INSEE): As part of the previous project, JDemetra+ underwent a major overhaul with the development of version 3.x, which offers algorithm extensions, such as seasonal adjustment of high-frequency data. The current initiative will aim to consolidate and finalise this new version, by further extending certain algorithms, improving the documentation, making the R packages available on CRAN, and adapting JDemetra+ to new computing environments, such as cloud infrastructure.