Exploring the thematic landscape of the circular economy. A comparative study of publication and Wikipedia data

The transition towards a sustainable society is one of the most relevant topics in recent studies of all sciences. Here, the thematic landscape around the topic circular economy (CE) is of great importance for researchers and policy actors. In our approach, we strive to represent this landscape based on a publication and Wikipedia dataset. Concerning the first, we retrieve scientific publications related to CE from the Dimensions database with the help of a suitable search query. With regard to the latter, we build on the hyperlinks of the Wikipedia page of CE to build up a database of CE related Wikipedia pages. We cluster both datasets separately with an approach that combines BERT embeddings and noun phrases of the publications' abstracts and the page content of each Wikipedia page. Our preliminary results based on manual comparison and evaluation of the cluster results of both datasets show that for the case of CE publication data and Wikipedia data match each other on established topics. Moreover, there are topics to be found in both datasets that are unique, which may represent niche, declining or emerging topics. This methodology attempts to contribute to the scientific discourse by 1) evaluating the potential of Wikipedia for bibliometric analyses compared to publication data, 2) illustrating the thematic landscape of both data resources with the example of circular economy and 3) identifying emerging topics or recent trends that are not commonly represented, highlighting the complementary power of Wikipedia data for bibliometric studies.