The nature of scientific datasets in South American repositories: a survey of formats and extensions

Abstract

Objective: identifying the scientific data repositories created and managed by Higher Education Institutions and/or South American research and funding agencies; identifying and describing the formats and extensions of files that compile the scientific datasets deposited in these repositories.

Methods: eight repositories retrieved by RE3DATA were selected for investigation. A population (N) of 1.115 scientific datasets was obtained. By using Stratified Random Sampling, the resulting sample (n) value was 258 datasets, which corresponds to 23,15% of the population (N). Data surveyed from the samples were condensed into tables and charts.

Results: it was noticed that the nature of the scientific datasets investigated is centered on textual and numerical data, saved in text files and tables, respectively. Also, the datasets may be either homogeneous (one or more files saved in a unique format and extension, e.g.: image format in .jpg) or heterogeneous (files saved in different formats and extensions, e.g.: same image format saved in .jpg and .tiff) in their composition. It was found that some extensions enable the identification of the nature, domain and content of the data, as observed in the .gpx and gdb extensions, which refer to geospatial data, therefore, alphanumeric data.

Conclusions: There is a growing need of describing the nature of data, as well as the formats and extensions of files. This kind of descriptive metadata would be valuable to potential users, as it would allow a greater understanding of the context of the data, focusing on data reuse.