« Some time ago, we at INSEE developed the habit of presenting our goals in a strategic document. INSEE Horizon 2025 is therefore pursuing this long-established tradition. Although the purpose of this type of exercise is certainly not to try to describe the future with any accuracy, and although INSEE is a sound and recognised institution, we are nevertheless sure that it is essential to occasionally take time for a moment of strategic reflection. The aim of INSEE – the aim which unites and motivates all of its employees – is to continue to inform the economic and social debate, as it has done for 70 years. In this document INSEE presents its ambitions for the future. »
Can statistics and Big Data tackle unemployment in Europe? From 13 to 15 March, participants at the European Big Data Hackathon organised by Eurostat in Brussels were set the task of combining the use of new types of data, such as data scraped from internet, with traditional statistical sources, such as for example data from European statistical surveys on employment (Labour Force Survey), income and living conditions (SILC) and adult skills (PIAAC).
The INSEE team was ranked 2nd out of 22 teams representing their European National Statistical Institutes competing in the Big Data Hackathon . The winner was Croatia and Estonia finished third.
In just two days the teams had to come up with prototypes that could answer the question “How can we use data to help EU decision-makers reduce the mismatch between job offers and skills available?” In some regions, jobs requiring certain types of skill find no takers, while in other areas, people who have these skills are looking for work. The aim of the competition was to develop prototypes using different data sources and to provide a better understanding of the problem, especially through data visualisation tools.
The French team from INSEE rose to the challenge. Team members were Yves-Laurent Bénichou, IT engineer and innovation project leader, Stéphanie Combes and Benjamin Sakarovitch, both data scientists in the Statistical Methods Department. Using data taken from the EURES website which collects CVs and job offers at European level, the Labour Force Survey, the ESCO classification of skills and occupations and scraped data on job offers, then applying their expertise in machine learning, data visualisation and data-architecture, the French team was able to present an interesting and promising prototype. Entitled “M&M’s”, short for “Migrations & Mismatch of Skills”, it offers a view of numerous phenomena, such as “Which skills are under-exploited in such-and-such a region while qualified people remain unemployed?”
Projects were evaluated by a 20-member panel made up of decision-makers from several directorates and divisions and industry representatives from Oracle, Amazon, Microsoft, IBM, SAP, SAS, Accenture, etc. Marianne Thyssen, European Commissioner for employment, social affairs, skills and labour mobility, welcomed the result of this experiment.
“This Big Data Hackathon organised by Eurostat shows how committed the various European statistical institutes are to exploiting the Big Data which our digital society continues to generate and also the related statistical processing techniques. I am very proud of the work our team has achieved. It is part of the ongoing process of improving our methods to mobilise intelligently, make our statistical production more meaningful and incorporate new data sources”, said Sylvie Lagarde, Director of the Methodology, Statistical Coordination, International Relations Directorate at INSEE.