Open Conference Systems, STATISTICS AND DATA SCIENCE: NEW CHALLENGES, NEW GENERATIONS

Font Size: 
Emerging challenges in official statistics: new sources, methods and skills
Giorgio Alleva

Last modified: 2017-05-22

Abstract


Official statistics is challenged to provide an increasingly complete picture of the complexity of our societies with a compelling demand for data. At the same time, it is facing human resources and budget constraints.

The development and dissemination of new digital technologies have removed many obstacles, first of all the cost for production, storage and analysis of information. A leap forward in efficiency is therefore in order, if we are to meet our responsibilities and to guarantee ever increasing quality standards.

As sampling surveys are expensive, response rates are decreasing and response burden must be reduced, data collection need to be optimised. The emergence of new data sources and availability of Big data and the opportunity of a massive exploitation of those already at hand (like administrative data) require new tools and methodologies.

The response to these thematic, methodological and organizational challenges lies in “integration”: of sources, of methods, and of skills.

Multiple use of data sources should be based on a re-engineering of the production process of official statistics. At Istat, the core of the new organisation aims at moving away from the ‘silo’ approach, typical of traditional statistical agencies, towards the enhancement of horizontal services: management, methodology and IT innovations drive the integration process, linking sources to boost coherence, tailoring new products to the different users’ needs, reducing the response burden through the reuse of available data and information, increasing the use of technology, and resulting in significant efficiency and time saving.

This new organisational model supports the Integrated System of Statistical Registers, a single logical data asset resulting from the integration of survey data, administrative data as well as data coming from new sources. Pillars of this system are the Population Register, the Business Register and the Territorial Register which are interconnected with one another through the Activity Register. The Integrated System allows achieving units and variables identification and estimation consistency as single cohesive units, which will make several new analyses possible (including a longitudinal approach). Hence, the system will not only improve efficiency by means of economies of scale, but also high quality and richer statistical outputs.

The path towards integration cannot leave the research activity of the Institute aside: research and development of new techniques and methodologies are indeed at the center of Istat’s modernisation project. Istat has set up a three-year plan for methodological and thematic research. Methodological research will move along four strategic research areas: integrated system of registers; censuses obtained by data integration; big data; and the unique process. Innovation will emerge from the so-called “Innovation lab”, a place where researchers will share new ideas, test new solutions, new processes and new products.

The frontier of data integration for official statistics is represented by the increasing opportunities to use big data to produce timely high-quality statistics with greater detail, and competing with growing numbers of new, non-official, players. NSIs are compelled to speed their production and make it more effective and less burdensome for respondents.

In Istat, as in most NSIs in Europe, several projects using big data sources for the production of statistics are currently ongoing. Some projects are in the early stages of implementation, some other are still in the experimental phase. We expect to reap the first results in the near future. All those projects need to tackle three key issues: quality, privacy and security issues, partnerships.

The use of big data in the production of official statistics is part of a wider strategy on “Experimental statistics” including new indicators from integration of sources, new tools for new phenomena, and unconventional classifications. Outputs from these innovative work will need to be treated accordingly.

Data dissemination is another key issue in official statistics’ innovation with the progressive opening of our data at its core. Open data are a key enabler of data driven innovation. When official statistics meets open data, several benefits are generated: from the possibility to reach users more easily to the enrichment of the published information with metadata that allow a proper interpretation. Much has already been done in the last years to disseminate them. Through the Linked Open Data portal users can now access interconnected and structured information through graphical interfaces that can be directly queried by external applications, independently of the technologies adopted.

Finally, on the way towards innovation, high level skills and a change-driven culture are essential.

Statistical institutions need data specialist able to produce, integrate and interpret data and to work with big and open data, but such skills are also strongly requested by the market, everywhere in Europe. Italy needs to engage further to urgently foster these new professions. Statistical organisations can also benefit greatly from each other and from mutual exchange and support and networking.