This dataset contains 404 mulberry disasters found in a digital collection of 4,000 Chinese local gazetteers (published in Erudition's Zhongguo Fangzhi Ku, or the Database of Chinese Local Gazetteers) that was curated during 2018 and 2019 by scholars at the Max Planck Institute for the History of Science to support their joint research paper entitled: “What Is Local Knowledge: Digital Humanities and Yuan Dynasty Disasters in Imperial China’s Local Gazetteers.” The paper appeared in the Journal of Chinese History in its 2020 spring issue. The authors use this dataset to ask what results the emerging methodology of analyzing data drawn from historical sources can produce for historical research, given that such data-driven analysis is inevitably quantitative and that many historians believe this would contradict the core value of studies in the humanities. This dataset is accompanied by a data paper to be published by Brill's Digital Concordances Platform. In the data paper, the authors give information about the full curation cycle of this dataset, including the research goal that motivated the curation, the sources that were used, and the curation/cleansing/preparation process. The authors also explain this dataset in detail, including the kind of information it contains and its overall temporal and geospatial distributions. In conclusion, they discuss the potential usages of the dataset. With this paper, we also hope to offer an exemplary data curation workflow for collecting data from historical sources that includes not only data curation (whether manual or semiautomatic), but also the practical steps of data cleansing and normalization that reflect the various historical considerations in the process.