Data science tools and techniques in the central bank environment have mainly focused on policy-related applications. For example, machine learning techniques have been used to develop high-frequency indicators of inflation, and the measurement of people’s inflation perception using social media data sources, etc. In parallel, the exponential rise of available structured and unstructured data in the internet allows statistical organizations to build, enhance or collect data and indicators in new ways. In this new digital age, indicators stemming from large data sets of multinational firms can be constructed in a timelier manner, with the help of public-private partnership and other types of co-operation. A prerequisite for those indicators are sound methodologies and a standardized exchange of data in suitable data formats. This development is at the research frontier, and there are encouraging examples in Banking Supervision, Fintech Monitoring and information extraction for Sustainable Finance purposes. The approach promises real time information that comes without the contorted process of enacting mandatory reporting.

The aim of this session is to focus specifically on how data science can contribute (and have contributed) to the statistics production process at central banks. Some topics include (but not limited to): (1) the automation and optimisation of statistics production processes (including dissemination and reporting), (2) the use of novel techniques in the areas of cluster analysis, outlier identification and treatment, as well as survey-level imputation, (3)  the development of open-source data management platforms and its integration with existing internal systems , (4) internally-developed business intelligence (front-end) solutions for internal and external stakeholders, and (5) the sourcing, processing and dissemination of statistics using non-traditional sources of data. Lastly, central bank representatives will showcase what has been done in their respective areas on the topic. There is in particular interest in the toolchains and workflows used by central bank data scientists to enhance the sourcing, compilation, and dissemination process. The interactions with the information technology department regarding architectural designs (tech stack), the introduction of new technology, and cybersecurity considerations (and the associated challenges/opportunities) would also be analysed.

Organiser: Mr Bruno Tissot 

Chair: Ms Gloria Peña 

Speaker: Mr Martín Rebolledo 

Speaker: Mr Alejandro Morales 

Speaker: Mrs Anggraini Widjanarti 

Speaker: Dr Andrea Del Monaco