Topic-emotion co-evolution in sixteen years of Wuhan rainstorm discourse on Weibo (2010-2025): topic assignments and aggregated tables

<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="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>