2026-10-08T07:46:29Z https://api.figshare.com/v2/oai

oai:figshare.com:article/33960343 2026-09-22T00:28:02Z category_28849 category_25474 category_29401 item_type_3 month_year_09_2026

Topic-emotion co-evolution in sixteen years of Wuhan rainstorm discourse on Weibo (2010-2025): topic assignments and aggregated tables Si Wang (8957504) Natural language processing Disaster and emergency management Corpus linguistics topic modeling BERTopic sentiment analysis social media Weibo disaster informatics urban flooding sarcasm risk communication Wuhan rainstorm Chinese NLP <p>Topic assignments, topic taxonomy and aggregated topic x emotion tables of a study of topic-emotion co-evolution in sixteen years of Wuhan rainstorm discourse on Weibo (2010-2025). This record builds on the WuRain-Senti corpus record (<a href="https://doi.org/10.6084/m9.figshare.33302052">https://doi.org/10.6084/m9.figshare.33302052</a>), which carries the relevance and emotion labels themselves; <code>post_id</code> is the same salted hash in both records, so they join post by post without exposing any identifier.</p> <p><b>Contents.</b> BERTopic assignments for all 24,537 flood-relevant posts (with a flag for the dedicated "Wuhan sea-viewing" retrieval channel); the 45-topic taxonomy with English labels and seven macro-groups; the ten class-based TF-IDF terms of each topic; the topic x emotion profile of the 5,910-post intersection; yearly counts of the seven groups; the emotion composition of the 10,942-post emotion tier by period, flood season and benchmark event month; the period composition of the annotation pool; the monthly ERA5-Land rainfall series (192 months, Beijing time) that gives the benchmark event months their rainfall context; and <code>code/recompute_paper2.py</code>, which recomputes every checked value of the paper from these files plus <code>labels_10942.csv</code> of the WuRain-Senti record and ends with ALL CHECKED VALUES MATCH THE PAPER.</p> <p><b>Not included.</b> Post texts, user identifiers, timestamps finer than the calendar year, representative documents, the BERTopic pipeline, the classifier, figure scripts and the crawler.</p> <p>Data under CC BY 4.0, code under MIT. ERA5-Land is derived from Copernicus Climate Change Service data. Please credit both this record and the WuRain-Senti record when reusing.</p> 2026-09-22T00:28:02Z Dataset Dataset 10.6084/m9.figshare.33960343.v1 https://figshare.com/articles/dataset/Topic-emotion_co-evolution_in_sixteen_years_of_Wuhan_rainstorm_discourse_on_Weibo_2010-2025_topic_assignments_and_aggregated_tables/33960343 CC BY 4.0