News As Digital Data: Text Mining and Analysis of Online News With Knime


Ümit Atabek, Gülseren Şendur Atabek




This paper considers news as a text corpus. The computational analysis of the news from this perspective requires data mining and analysis techniques that can handle large amounts of unstructured dataprocessing. KNIME, a free andopen-source software (FOSS) that is developed for data science applications, is very suitable for news analysis from this perspective. KNIME offers advanced text mining and analytics capabilities. Here, we develop an example workflow for online news content analysis, as well as a text-network analysis. Furthermore, we explicatein detail the workflow and the KNIME nodes used for these analyses. Our proposed workflow is reasonably versatileand flexible to be applied to other journalistic textual analyses such as similarity, sentiment, frame, discourse, and thematic analyses. This workflow also exemplifies how no/low codedata processing and computing could be effectivelyemployed in journalism and media studies.