10.7910/DVN/YJVIOQ Gu, Hengyu Hengyu Gu https://orcid.org/0000-0002-1174-4940 Nanjing university Wu, Yingju Yingju Wu https://orcid.org/0000-0002-9735-0992 Nanjing university Marois, Guillaume Guillaume Marois https://orcid.org/0000-0002-2701-6286 Shanghai university Lutz, Wolfgang Wolfgang Lutz https://orcid.org/0000-0001-7975-8145 Shanghai university Niu, Tianlong Tianlong Niu https://orcid.org/0009-0006-6166-8997 Nanjing University Harvard Dataverse 2026 Social Sciences Age support ratio Task-based Skill Ratio Wu, Yingju Yingju Wu Nanjing university 2026-02-02 2026-02-04 2026-03-25 14143 2521153 1386191 application/vnd.openxmlformats-officedocument.wordprocessingml.document application/zip application/zip 2.3 Creative Commons CC0 1.0 Universal Public Domain Dedication. This repository provides a comprehensive replication package supporting the findings of this study, comprising the city-level panel data and Stata codes necessary to reproduce the main empirical results (Table 1), along with the underlying datasets for the spatiotemporal visualizations (Figures 1 and 2) and projection simulations (Figure 5). Please note that the datasets provided herein also serve as the source for all remaining figures and visualizations beyond those explicitly listed. To ensure transparency and reproducibility in constructing the Task-based Skill Ratio (TSR), we provide intermediate data covering the entire process from occupational description matching to task attribute classification, including multi-source classification results from manual coding, GPT-5, and Gemini 2.5 Pro for robustness verification.