Why is this the case? What academic research is being cited and who is citing it? What motivates Stack Overflow users to refer to academic research? How do they use academic research and what are the implications of this usage? ......
To shed light on these questions, we presented what we believe to be the first large-scale collection and analysis of academic references on Stack Overflow. Our study sifted through 44 million URLs found in the edit histories of 59 million posts (totaling 160 million revisions) and identified 15,066 references to 10,718 unique academic articles published in over 2,900 venues. We have made the dataset available for download, aiming to facilitate further research into the interaction of academic knowledge and discussions about practical challenges on one of the largest technical forums online. This website interactively presents the findings of our study.What Fields of Research (FoR) are referenced in different discussions?
Choose a technical domain to see which Fields of Research are cited within the discussions of this domain.(Note: FoRs not being cited are not shown)
Choose a Field of Research to see how it is referenced across different technical domains of Stack Overflow discussions.
How do Stack Overflow communities interact with academic research?
To better interpret and visualize the interaction between various technical domains and academic research, we first simplified the bipartite network by aggregating 60+ Fields of Research into 7 broad disciplines. The heatmap below illustrates the distribution of references to different Fields of Research across 16 technical domains. Cell (x,y) denotes the percentage of papers referenced by domain x that originate from discipline y. For example, 79.9% of the articles referred in domain D0 are from the AI/ML/CV/NLP discipline.
We further map the structure of scientific knowledge on Stack Overflow through a co-citation network below. Each node in this network represents an academic article and is connected to another if they were jointly referenced by the same post. The node size reflects the article’s PageRank score, denoting its importance and influence across the network