Curated Dataset of Celebrity Deaths with Cardiovascular and Non-Cardiovascular Classification for Exploratory Mortality Analysis (n=164)

This dataset contains 164 curated records of publicly reported celebrity deaths, with demographic variables, reported cause of death, and classification into cardiovascular and non-cardiovascular categories. In version 1.1, cardiovascular cases were further subdivided into six ICD-10–aligned subcategories.

Cases were identified through publicly available English-language sources, principally Wikipedia entries and their cited references. The dataset is curated and is not population-representative; it is intended for descriptive and methodological analysis rather than epidemiological inference.

Version 1.2 adds five supplementary methodological files: a case-level ICD mapping dictionary, a phrase-level mapping summary, a retrospective sampling-log documentation file, an inter-rater reliability record, and a reproducible Python analysis pipeline. The 164 case records are unchanged from version 1.1.

This dataset underpins the study: "Patterns in Curated Celebrity Mortality Data: An Exploratory Analysis of Cardiovascular Disease Representation and Dataset Bias."