Big Data for Macroeconomic Statistics

Source details

Intended participants:
- Technical staff and managers from national statistics offices and central banks involved in macroeconomic statistics, nowcasting, and data science projects.
- Economists, statisticians, data scientists, and decision-makers overseeing related projects or teams.
Qualifications:
- Expected to have a degree in economics, statistics, data science or equivalent experience.
- Working knowledge of advanced programming languages such as Python and R is strongly recommended.

Audience: Country officials from national statistics offices (NSOs) and central banks seeking skills to leverage Big Data and advanced technologies for high-frequency and detailed statistics.
Delivery: Lectures and hands-on workshops delivered by data scientists of the IMF Big Data Center.
Emphases:
- Big Data applications pertinent to macroeconomic statistics from both producer and user standpoints.
- Participant presentations and discussion of use cases, peer learning, and experience sharing.

Technical architecture and workflows:
- Provide support in setting up Big Data architecture, encompassing data extraction, preprocessing, and visualization in their various organizations.
Machine learning and nowcasting:
- Demonstrate a practical knowledge of machine learning modeling, variable selection, performance analysis as well as model selection, for GDP nowcasting and trade monitoring (based on the IMF's "PortWatch" platform).
Geospatial analysis:
- Demonstrate practical skills in the use of Google Earth Engine (GEE), Dynamic World, Jupyter Notebook and other geospatial and data science packages to analyze satellite data and generate high-frequency macroeconomic statistics.
Textual analysis and NLP:
- Carry out textual analysis with natural language processing (NLP) technologies to support macroeconomic analysis.
Statistical production improvements:
- Demonstrate the application of these Big Data technologies and resources to improve timelines and granularity of their official statistics.
Institutional capacity and collaboration:
- Facilitate peer-learning on Big Data applications and explore collaborations between agencies working on projects of mutual interest.
- Gain insights into setting up effective data science teams and developing institutional Big Data strategies to support innovation for macroeconomic statistics.