The Semantic Networks from news articles (German sample) contains semantic networks for a sample of migration-related news articles extracted from the dataset described in Pluralistic Recommendation in News [1]). The networks, i.e., nodes and edges, are generated using the Emoatlas library [2] on the raw news texts in German. Each news article is identified by a random UUID, i.e., uuid.uuid4(), and is represented by the unweighted edges and nodes of the network. In addition, it is enriched with the publication date, the name of the media (in anonymized form), the country, and the language. [1] Pluralistic Recommendation in News, url https://data.d4science.net/TDpx. [2] Semeraro, A., Vilella, S., Mohammad, S., Ruffo, G., & Stella, M. (2023). EmoAtlas: An emotional profiling tool merging psychological lexicons, artificial intelligence and network science.