Replication Data for: Where the wind blows: Five Star Movement’s populism, direct democracy and ideological flexibility

This dataset includes the corpus of online articles on which the paper is based. In detail, the files provided are: 1) the full blog data that was downloaded and structured (BG_BLOG_till2015_11_wTOPICS.enum.xml) 2) the full blog data that was downloaded and structured; additionally, all the texts are parsed and the parses are saved in the XML file. (BG_BLOG_till2015_11_wTOPICS.enum_spacy_parsed.xml) 3) a list of labels (categories) and corpus_IDs which refer to the content labelled as being an example of this category. On this data we train the nearest centroid classifier. (labels_7cat_articles_numbers_only.txt) 4) same as the list before but with less labelled data (96 articles; like in the paper) (labels_7cat_articles_numbers_only_reduced96.txt) 5) a python script which replicates the results for the evaluation of the nearest centroid classifier, using 5-fold cross-validation and addi- tionally a leave-one-out evaluation for accuracy as well. (nearest_centroid_eval.py