Twenty Years of Building Simulation Trends: Text Mining and Topic Modeling of the Bldg-Sim Email List Archive

Abstract

The recent growth of the building performance research community has been in parallel with the one building.org Bldg-sim email list. This list was formed in 1999 and steadily grew into a major venue for building simulation community announcements, discussions, and questions and answers (Q&A). This paper presents an analysis of the Bldg-sim email archive to determine how traffic has grown over the years and what general trends have come and gone in this particular user community. We use a text mining approach to find the most prominent topics over the years and create trend metrics from the frequency of the most common words from those topics. The results illustrate the relative rise and fall of various software tools, organizations, and simulation topics from the portion of the simulation community who used the email list. Visualizations showing the trends are presented and discussed. The paper also discusses the generalizability of results as it should be noted that the users of this email list are primarily English speaking simulation practitioners from North America. All data and code used in this analysis are available in a reproducible GitHub repository.