Datasets for: Declining Modularity of Intellectual Bases During the Emergence of Research Areas

This deposit contains the derived datasets that accompany the paper:"Kazuki Nakajima, Yuya Sasaki, and Masaki Aida. Declining Modularity of Intellectual Bases During the Emergence of Research Areas (2026)." The paper proposes a framework that tracks the modularity of co-citation networks over sliding time windows, evaluates its changes as bootstrap distributions with effect sizes, and identifies papers strongly associated with the decline. The framework is applied to three research areas: higher-order network science (OpenAlex), superstring theory (INSPIRE HEP), and graph representation learning (OpenAlex). Contents data.zip: deduplicated source/target paper corpora and derived inputs for the three research areasresults.zip: bootstrap checkpoints (B = 1,000 per time window) and derived statistics underlying the paper's figures and tablessearch_keywords.xlsx: all keyword queries used to define each research areaREADME.md: full description of the file structure, reproduction tiers, and usage notes Together with the analysis code at https://github.com/kazuibasou/cocitation-modularity (MIT license), these files reproduce the paper's numerical results, tables, and data-driven figure panels. See README.md for the exact scope and instructions. Bibliographic records are derived from the OpenAlex public snapshot of September 30, 2025, and the INSPIRE HEP snapshot of January 8, 2021.