Granger causality between usage counts and publication numbers

In this paper, 3,975,602 articles from IEEE Xplore and their 313,099,560 usage data points generated during a 120-month period from January 2011 to December 2020 were used as the study sample, covering 64,633 topics in total. The temporal relationships between the usage counts and publication numbers of these topics were examined via Granger causality analysis. The empirical findings show that nearly 80% of the topics exhibit significant usage-publication interactions from a time-series perspective. On the other hand, newly-emerging topics present more likely bidirectional usage-publication interactions, but for more mature topics, publication numbers tend to affect usage counts only unidirectionally.