The PREDICT Dataset Methodology includes three main sections and an Annex:
- The Core PREDICT Dataset: describes the content of the dataset and its construction process.
- The Backwards Reconstruction of Time Series: describes the reconstruction of the series until 1995, detailing the sources used when data are not available at the level of detail needed.
- Nowcasting Methodology: describes the methodology followed to nowcast the main variables for two additional years. A sub-section addresses the methodological issues arising with nowcasting in times of economic crisis, like the one caused by the COVID-19 pandemic.
- Annex I: presents correspondence tables between different industrial classifications, and definitions of ICT occupations employed in PREDICT.
Further relevant information related to the data collection, including references to official sources of raw data used to calculate the PREDICT statistics for each country, can be consulted in the 2022 PREDICT Dataset Methodology