Eurozone Monetary Policy Transcripts & Metadata

Columns

• speech_id: A unique numeric identifier assigned to each row/record.
• when_speech: The date the item was released or delivered (Format: YYYY-MM-DD).
• who: The name of the speaker or participant (e.g., Jean-Claude Trichet, Benoit Coure), with accents removed for easier searching.
• what_title: The official title of the speech, press conference, or event (e.g., "Introductory statement with Q&A").
• what_frequencies: A JSON-encoded string containing the frequency count of words within the content, enabling the generation of tag clouds (e.g., {"crisis":12, "financial":10...}).
• what_language: The 2-character ISO code indicating the language of the content (predominantly "EN" for English).
• what_weblink: The direct URL linking to the original source material on the ECB website.
• what_type: A single-character code categorising the item: 'S' (Speeches), 'P' (Press Conferences), 'I' (Interviews), 'B' (Blog posts), or 'E' (ECB Podcasts).

Distribution

• Format: CSV (Comma Separated Values).
• Size: Approximately 6.59 MB.
• Structure: 8 columns; approximately 4,171 valid records.
• Update Frequency: Weekly (typically Monday evenings).
• Note: Includes AI-based audio transcripts for accessible audio files since March 2024.

Usage

• Sentiment Analysis: Evaluate the shifting tone of central bank officials regarding inflation and market stability.
• Topic Modelling: Identify recurring themes and keywords (e.g., "crisis", "monetary policy") using the pre-calculated word frequencies.
• Historical Research: Trace the evolution of ECB communication strategies from 1997 to the present.
• Search Engine Development: Utilise the tagged and structured data to build custom search applications or "search by association" tools.
• NLP Training: Train Natural Language Processing models on specific financial and economic vocabulary.

Coverage

• Geographic Scope: Eurozone / European Union.
• Time Range: 7 February 1997 to present (dataset includes records up to November 2025, likely covering scheduled events or projections).
• Demographic: primarily high-level ECB officials, board members, and presidents.
• Content Types: Initially speeches and interviews; expanded in 2020 to include press conferences, blog posts, and podcasts.
• Language: Primarily English (96%), with minor instances of German and other languages.

License

Who Can Use It

• Economists: For analysing monetary policy trends.
• Data Scientists: For NLP projects and visualisation.
• Financial Analysts: To correlate bank communication with market movements.
• Journalists: For fact-checking and referencing historical statements.
• Students: For academic research on European economics.

Dataset Name Suggestions

• ECB Communications & Speeches Archive (1997-Present)
• Eurozone Monetary Policy Transcripts & Metadata
• European Central Bank Searchable Text Corpus
• ECB Speeches, Interviews, and Press Conferences Data

Attributes