COVID-19 Media Analysis: Lemma and Entity Frequencies for Switzerland, Jan-Jul 2020

A multilingual corpus of news articles regarding Switzerland and COVID-19 was analyzed for lemma and named entity frequencies. The dataset was created by Giovanni Spitale at the University of Zurich using a custom NLP pipeline on articles from Factiva's major global outlets. Articles were published between January and July 2020 and were filtered to be over 300 words and in German, French, Italian, or English.

Use Cases

• Track media attention to specific entities like organizations or locations based on named entity frequencies.
• Analyze linguistic trends and topic evolution over time based on lemma frequency data.
• Compare COVID-19 media narratives across German, French, Italian, and English languages.
• Study the information environment in Switzerland by analyzing articles from both domestic and foreign media.

Strengths

• Analysis is based on articles from major global news outlets accessed via Factiva.
• Data covers a defined 7-month period from January to July 2020.
• Multilingual coverage includes German, French, Italian, and English articles.
• Query design targeted articles over 300 words and specifically about Switzerland.

Limitations

• The original article text is not shared due to copyright restrictions; only analysis results are provided.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

Factiva (Dow Jones) news database.

Collection Method

Articles gathered via targeted Factiva query and processed with an ad-hoc parser and NLP pipeline.

Time Range

January 2020 to July 2020

Geography

Switzerland (as the subject of articles, not the origin of media)