Project Spectrum: using generative AI to enhance inflation nowcasting

Project Spectrum – a collaboration between the Bank for International Settlements (BIS), the Deutsche Bundesbank and the European Central Bank – has successfully combined text embeddings with machine learning algorithms to efficiently classify millions of individual products for inflation analysis. The method was tested on the European Central Bank's Daily Price Dataset (DPD), which contains billions of price-product daily observations for 34 million unique products.

By turning raw, fragmented product description into structured data, Project Spectrum equips analysts and policymakers with timely, detailed insights into price developments. Ultimately, it contributes to an emerging field of AI-powered analysis for more accurate economic modeling and forecasting.

This report is intended for monetary policy analysts who utilise high-frequency data for inflation nowcasting and data scientists within central banks looking for cost-effective alternatives to large language models for large-scale classification. It also serves as a technical reference for statistical agencies seeking to automate the categorization of scanner and web-scraped data into official indices. Finally, it provides a methodological framework for economic researchers studying price-setting behaviour at the individual product level.