Nowcasting (NWC)
Deadline passed
Session No.: AT 23.03
Location: Ebene, Mauritius
Date: January 23-27, 2023 (1 week) New dates
Delivery Method: In-person Training
Primary Language: English
Interpretation Language: French Portuguese
Target Audience
Junior and middle-level officials from Ministries of Finance, Central Banks, and other interested public institutions.
Qualifications
Basic knowledge in statistics and statistical packages such as R and/or Eviews. The course will aim at using R but participants who have access and knowledge of EViews will receive adapted materials.
Course Description
The course will introduce participants to nowcasting methods and to the type of data and techniques that countries can use to develop high-frequency indicators (HFI) using a hands-on approach. The course will cover nowcasting and HFI data sources, two types of modeling approaches (factor models and machine learning), that enable the incorporation of mixed-frequency data indicators to the nowcasts and forecasts of GDP. Using applications to different countries, it will emphasize the following practical aspects: How to adapt to the data context? How to deal with mixed frequencies and delays in the release of data? How to evaluate (and combine) different nowcasting models?
Course Objectives
Upon completion of this course, participants should be able to:
- Hone officials; skills on nowcasting techniques and high-frequency data concepts, illustrating the use of different types of data (e.g., webscraped data, APIs, crowdsourced data, scanner data) .
- Build institutions; nowcasting capacity.
- Enhance forecasting quality to support policymaking.
- Encourage technical exchange on different methodologies of assessments and facilitate cross country collaboration.
- Integrate these new elements into a wider framework of technical assistance, such as the monthly indicator of economic activity (STA), and work on data compilation and dissemination.
- Contribute to integration of capacity development and policy engagement and program work