Corpus-driven conceptual analysis of epidemic and coronavirus for the Humanitarian Encyclopedia

Abstract

Terminological conceptual analysis can be applied to purposes beyond terminology work. This article presents a Frame-based Terminology approach adapted to analyse concepts and inform the content of entries in the Humanitarian Encyclopedia. It proposes a method for conceptual analysis by systematising the extraction of knowledge rich contexts (KRCs) around corpus querying tasks through semantic sketch grammars (SSGs) and macros with knowledge patterns (KPs). KRCs are curated manually, modelled into conceptual propositions, and combined with corpus metadata into unified datasets. The method was tested on epidemic and coronavirus and their results are presented. This study provides a preliminary model to operationalise the study of conceptual variation. It also identifies the areas of terminological conceptual analysis with the potential to be informed by other research methods towards creating a standalone methodology.

Quartz: A Template for Quantitative Corpus Data Visualization Tools

“…A more detailed examination of concordances could be orchestrated by refining a series of cross-filters. This approach would enhance the feasibility of conducting a more complete conceptual analysis (Chambó and León-Araúz 2023a) by making the data more manageable, since the qualitative analysis of tens of thousands of occurrences would be infeasible. Advanced visualizations can simplify the observation of variations and commonalities that would be outside of the scope of a single CMS window (i.e., requiring more operations for side-by-side comparison).…”

“…It constitutes a domainspecific terminological resource serving experts and the wider humanitarian community. The encyclopedia's content is primarily derived from corpus-based methods utilizing SkE, with visualization taking an important role (Chambó and León-Araúz 2021). Visualization methods are used both to explore data and present results, and beyond being a feature of the encyclopedia, they play a role in operationalizing the study of conceptual variation, in this case for the humanitarian domain (Chambó and León-Araúz 2023b).…”

“…The visualization of data for HE has expanded since its inception, from utilizing default SkE features (Figure 1) to a range of chart types built with third-party software, including the Tableau visual analytics platform (Figure 2), Flourish, and the ggplot package in R. SkE's bar charts for text type frequencies are more rudimentary but are generated automatically and integrated with the query system (nearby buttons allow navigating to concordances for a data sample). The Tableau charts, which have been described in detail in previous work (Chambó and León-Araúz 2021), were designed to increase interactivity: related charts are accessible via tabs; data points are selectable, hoverable, and sortable; and the presentation of data benefits from the various design capabilities of standalone visualization software. That said, several needs remained unmet while integrating visualization into the concept analysis procedure, needs often inherited from the more general challenges with data visualization in corpus linguistics, as enumerated in Section 1.…”