Page 403 - Contributed Paper Session (CPS) - Volume 4
P. 403
CPS2449 Louisa Nolan et al.
• identifying close-to-real-time big data and administrative datasets,
which represent useful economic concepts
• creating a set of indicators that allow early identification of large
economic changes
• providing insight into economic activity, at a level of timeliness and
granularity not currently possible with official economic statistics.
Although our indicators are not – and are not intended to be – a proxy for
GDP or other official statistics, they act as an early warning system for the UK
economy, addressing the need for faster economic information for decision-
makers. We currently publish them as monthly research outputs, and,
although these are not ‘official statistics’ they are regular, timely indicators,
which provide us with a new angle on economic activity, using data in a novel
way. We plan to further develop the shipping and road traffic indicators to
provide new indicators for the movement of goods into and around the UK.
The pilot project exploring the characteristics of high-growth firms has
produced some tantalising insights into how ‘non-traditional’ data can be
used to supplement our understanding of firm growth. We were able to
demonstrate how administrative, geospatial and textual data can be linked, at
a firm level, and used machine learning to give a richer insight into firm
behaviour than is possible with standard analyses of aggregate statistics.
Aspects of this work have now been adopted by the UK Department for
Business, Energy and Industrial Strategy (BEIS), to inform their policy
development.
The ability to take messy free text and translate it into syntactic and
semantic hierarchies has many potential applications, not only in economics.
In our project, we were able to summarise the goods being shipped by lorry
across ferry routes around the British Isles. It is unlikely that these will become
official statistics, as there is a wide variation in the quality of the data recorded
on the manifests, and there is no information on the volume and value of
goods being transported. However, in the absence of an existing survey or
access to more complete administrative data, the output of this project
delivered a rapid and valuable insight on the local movement of goods, where
previously very little information was available.
4.2 Conclusions
In this paper, we showcase some of the work of the UK’s Data Science
Campus. The projects discussed demonstrate how the combination of novel
data sources and the tools and techniques of data science can enhance the
evidence base for economics.
We learn that:
392 | I S I W S C 2 0 1 9