The dataset includes lexical networks centered on keywords related to migration. The networks are built starting from Finnish news articles extracted from the dataset described in Pluralistic Recommendation in News [1]). Each keyword is represented by a tuple (keyword_lemma, POS), e.g. for Finnish, migrant_NOUN, migrant_ADJ, emigrant_NOUN, emigrant_ADJ, izbjeglica_NOUN, izbjeglica_ADJ, imigrant_NOUN, imigrant_ADJ, migrirati_VERB. For each keyword_lemma the four previous and following words are extracted. The nodes of the networks are annotated with weighted attributes derived from sentiment and emotion lexicons, i.e., NRC lexicons. [1] Pluralistic Recommendation in News, url https://data.d4science.net/TDpx.