Co-citation Analysis

Published: 20 September 2019| Version 3 | DOI: 10.17632/4n8ns8vzvz.3
, James Bradley


The data in this version of the dataset are being actively annotated and supplemented. Please feel free to send email to the corresponding author for the Bradley et al. (2019) article, if you have questions. Access to the raw data used requires a Web of Science subscription that must be negotiated with Clarivate Analytics. Figure 1: Effect of Research Discipline, Background Network, and Citation Count on Conventionality and Novelty. Data are shown for the applied physics (18,305), immunology (21,917), metabolism (97,405) and WoS (476,288) networks for 1995; numbers in paren- theses are the count of publications in each network. Subfigures (a) and (b): the x-axis show publications classified into percentile groups based on citation counts (e.g., Top 1 in- dicates those publications in the top 1%) and the y-axis shows the percent of publications in each set that are HC or HN. Based on the selected background network, z-scores are computed for each disciplinary network; thus, imm denotes the immunology network with immunology z-scores and imm wos denotes the immunology network with z-scores from WoS z-scores.


Steps to reproduce

Custom ETL code for parsing WoS XML data into a PostgreSQL database are available at


Research Evaluation, Bibliometrics