Page 131 - Invited Paper Session (IPS) - Volume 2

IPS188 Bruno Tissot
2. Opportunities provided by Big Data
Given the challenges mentioned above, what are the opportunities offered
by big data – described by some as the new oil of the 21st century (The
Economist (2017))? The main sources of big data are social networks,
traditional business systems, and the internet of things, and four types of
3
data sets appear of particular importance in the economic and financial area:
internet-based indicators, commercial data sets, financial market indicators
and administrative records (IFC (2017)). These data are often available in an
4
“organic” way, unlike statistical surveys and censuses: the reason is that they
are usually not collected (“designed”) for a specific statistical purpose, being
the by-product of other activities (Groves (2011)). Hence, there is a clear
interest to use them, considering that the cost of launching traditional
statistical surveys can be significant. Moreover, the new type of information
they provide can help addressing some of the main challenges faced when
measuring and predicting inflation.

As regards the first issue related to the need for a robust statistical
infrastructure, big data can facilitate the work of public authorities
compiling inflation measures. It can certainly be an innovative source for the
current production of official statistics, offering access to a wider set of prices
(eg prices recorded online, credit card operations); it can also facilitate
statistical compilation work when the data are easier to collect compared to
traditional approaches. Indeed, data volumes have surged hand in hand with
the development of specific techniques for their analysis, with the emergence

3 Following the work conducted under the aegis of the United Nations (see Meeting of the
Expert Group on International Statistical Classifications (2015)).
Note: internet-based indicators is not necessarily the category the most used in this context;
4
see for instance the growing importance for public statisticians of the large datasets derived
from administrative records (Bean (2016)).
118 | I S I W S C 2 0 1 9