Abstract
This research introduces a comprehensive dataset of academic publications and professorial metrics from Iranian universities, systematically collected from Google Scholar using Python-based tools such as Selenium and BeautifulSoup, validated through expert review. articles.csv was kept raw except for exact duplicate removal, while a four-step Data Refinement Process (governmental affiliation, ≥ 100 citations, author-article verification, 2020-22 window) produced final_articles.csv for analysis. The dataset includes over 1.5 million records of articles scraped from various categories, providing detailed information on each article's title, citations, authorship details, and institutional affiliations, all curated through an intricate web scraping process. It spans multiple interlinked files with attributes including article metadata, professor profiles, and institutional details, We then applied a temporal filter (2020-2022) in conjunction with institution and author-level criteria, restricting to governmental universities and professors exceeding our citation threshold, and excluded records missing essential metadata (specifically, entries without titles or with removed/invalid Google Scholar links), yielding a focused cohort primed for downstream analytical pipelines. These attributes enable in-depth exploration of academic productivity, collaboration networks, and institutional performance across disciplines. The dataset provides a significant foundation for developing domain-specific models and theories related to academic impact and network analysis, with potential applications in social network analysis, trend identification in research disciplines, and benchmarking within the academic landscape. Additionally, it may support the development of machine learning and deep learning models for classifying research outputs and analysing scholarly trends, driving innovation in understanding academic ecosystems and informing data-driven strategies to enhance research excellence. This dataset is intended for use by scientometricians conducting citation and network analyses, university administrators performing institutional benchmarking, policymakers evaluating research strategies, and academic researchers exploring disciplinary trends. To remain fully compliant while capturing the complete corpus, we throttled requests far below Google Scholar's implicit limits and deliberately distributed the crawl over a three-month window, trading speed for a policy-conformant, loss-minimised harvest. In short, we rate-limited requests, avoided circumvention, and did not bypass access controls; we acknowledge Google's ToS restrict automated queries.