Big Data and Machine Learning for Macroeconomic Statistics
Source details
- Canonical URL
- Big Data and Machine Learning for Macroeconomic Statistics
Other formats
Bibliographic details
- Session: SA 25.23
- Location: New Delhi, India
- Dates: April 7-11, 2025 (1 week)
- Delivery method: In-person Training
- Primary language: English
- Status: Deadline passed
Program logistics and audience
- Session No.: SA 25.23
- Location: New Delhi, India
- Date: April 7-11, 2025 (1 week)
- Delivery Method: In-person Training
- Primary Language: English
- Status: Deadline passed
- Target Audience:
- Technical staff of national statistics offices (NSOs) and central banks (economists, statisticians, data analysts, or data scientists) working on macroeconomic statistics, nowcasting, and/or other data science projects.
Course description and content focus
- Organized and presented by the IMF Statistics Department for country officials from NSOs and central banks.
- Aimed to build knowledge and skills to leverage Big Data and Machine Learning for generating high-frequency and detailed statistics to enhance macroeconomic monitoring and research in the region.
- IMF experts will share real world applications based on cumulative experience of the IMF Big Data Center and the High Frequency Indicators for Economic Activity project.
- Program format:
- Lectures and hands-on workshops.
- Coverage of various Big Data applications and machine learning to enhance macroeconomic statistics and nowcasting.
- Insights from both producer and user standpoints.
- Participant presentations and discussion of use cases, peer-learning, and opportunities for inter-agency collaboration.
- Practical tools and platforms emphasized:
- Google Earth Engine (GEE)
- Jupyter Notebook
- Geospatial and data science packages
- IMF's "PortWatch" platform (trade monitoring example)
- Satellite imagery applications for GDP nowcasting
Course objectives
- Upon completion of the course, participants should be able to:
- ¿ Understand how big data and machine learning can complement traditional data and analytical techniques in macroeconomic analysis and nowcasting, through practical examples of GDP nowcasting (using satellite imagery) and trade monitoring (based on the IMF's "PortWatch" platform).
- ¿ Explore how these techniques may be used in the compilation of high frequency indicators.
- ¿ Demonstrate practical skills in the use of Google Earth Engine (GEE), Jupyter Notebook, as well as geospatial and data science packages to analyze satellite data and generate high-frequency macroeconomic statistics.
- ¿ Demonstrate the application of these Big Data technologies and resources to improve timelines and granularity of their official statistics.
- ¿ Participate in peer-learning on Big Data applications and explore collaborations between agencies working on projects of mutual interest.
Practical outcomes and expected benefits
- Capacity building in applying Big Data and Machine Learning methods to:
- Generate high-frequency macroeconomic indicators.
- Improve timeliness and granularity of official statistics.
- Support nowcasting efforts using satellite imagery and other high-frequency data sources.
- Strengthened cross-agency collaboration through participant use-case presentations and peer-learning.
References