Page 143 - Invited Paper Session (IPS) - Volume 1

IPS122 Elise C. et al.
The creation of this unit corresponds with the ESS vision 2020, which aims
to explore new opportunities of the digital transformation and build
organisations capable of working and collaborating with agility within the
official statistics service and with European peers (e.g. through the Big Data
task force and related ESSnets), but also with data producers and academics.
It is inspired by the experience of other NSIs in the Netherlands, Italy, Canada
and the U.-K. amongst others that have modernised their organisations to
respond to these challenges.

1.2 Context in France
The SSP Lab was created following experiments using Big Data sources
carried out in several of the INSEE and Ministerial Statistical Services units, e.g.
the project incorporating mass retail scan data for producing the CPI (INSEE,
started in 2011), the web scraping of job ads for estimating job vacancies
(Ministry of Labour), and the use of administrative health data for statistical
purposes (Ministry of Health). There has been also a positive regulatory
context for exploring new sources of data since the adoption of the so called
“Law for a Digital Republic” of 7 October 2016. At INSEE request, the latter
makes mandatory the transmission of information from internal databases for
companies concerned by a statistical survey. 1
The SSP Lab was also created following the restructuring strategy of INSEE.
INSEE set up in 2012 the Directorate of Methodology, Statistical Coordination
and International relations (DMCSI) with the idea of pooling rare and strategic
resources for possible synergies. Within the Department of Statistical Methods
(within the DMCSI), the Division of Applied Econometrics and Evaluation
(DMAEE) explored and disseminated innovative statistical and econometric
methods within the official statistics service by providing support and advice
to statisticians in charge of production. This division gradually integrated a
role of coordination and animation of the work on Big Data within the official
statistics service. It included two full-time data scientists in 2016. It joined the
network of European peers through the Big Data task force and the ESSnet Big
Data. It conducted several experimental projects on new sources, e.g.
evaluating the interest of Internet sources for nowcasting economic indicators
(Combes, Bortoli, Renault, 2015 and Combes, Bortoli, 2017), and started a
collaboration with the Orange SenSE laboratory on mobile phone data. In
response to requests from business units or Ministerial Statistical Services,
investments were also conducted to acquire skills and experience on textual
analysis and machine learning methods, and to disseminate practical
instructions. These investments provided the opportunity to launch reflections

1 It also cancels data transmission royalties between public administrations for
statistical purposes
132 | I S I W S C 2 0 1 9