Production of social statistics...goes social!
| Main Authors: | Grazzini, Jacopo, Lamarche, Pierre |
|---|---|
| Format: | Proceeding Journal |
| Terbitan: |
, 2017
|
| Subjects: | |
| Online Access: |
https://zenodo.org/record/3240501 |
Daftar Isi:
- The scope of this paper is to present a practical framework adopted for the integration of software applications into a statistical production chain. The focus is the actual implementation of a high-level collaborative platform aiming not only at producing social statistics, but also at further fostering experimentation and analysis in that field. In doing so, we strongly support the (obvious) claim that "the modernisation and industrialisation of official statistical production needs a unified combination of statistics and computer science in its very principles". Motivated by the consensus that processes - in particular statistical processes - for data-driven policy should be transparent, we naturally promote open, reproducible, reusable, verifiable, and collaborative software development and deployment in a statistical organisation. Beyond just devising guidelines and best practices, we show how the platform is implemented for the production of social statistics. For that purpose, we adopt a reasonable mix of bottom-up (from low-level scope to high-level vision) and top-down (from black-box process models to traceable functional modules) designs, so as to "think global, [and] act local". In building the parts while planning the whole, we provide with a flexible and agile approach to immediate needs and current legacy issues, as well as long-term problems and potential future requirements for statistical production.