Automatic Interpretation Biodiversity Spreadsheets Based On Recognition Of Construction Patterns

Abstract

Spreadsheets are widely adopted as "popular databases", where authors shape their solutions interactively. Although spreadsheets have characteristics that facilitate their adaptation by the author, they are not designed to integrate data across independent spreadsheets. In biology, we observed a significant amount of biodiversity data in spreadsheets treated as isolated entities with different tabular organizations, but with high potential for data articulation. In order to promote interoperability among these spreadsheets, we propose in this paper a technique based on pattern recognition of spreadsheets belonging to the biodiversity domain. It can be exploited to identify the spreadsheet in a higher level of abstraction - e.g., it is possible to identify the nature a spreadsheet as catalog or collection of specimen - improving the interoperability process. The paper details evidences of construction patterns of spreadsheets as well as proposes a semantic representation to them.