Descriptor Thesauri Renewable energy

This file allows the standardization of plural and singular terms, abbreviations or acronyms, as well as complete keywords, obtained from the information collected in Scopus in the categories of sustainability, environment, and renewable energies. The collected files were uploaded to the free software VOSviewer to calculate the number of occurrences of keywords. In addition, a co-occurrence threshold of ≥ 10 was set, a criterion that allowed the generation of a robust cognitive structure by ensuring the relevance of the selected terms. Consequently, 32,935 terms related to the ASJC 2105 code were retrieved, which, together with their frequency, were used to normalize the keywords and create a first thesaurus. VOSviewer generated a TXT file with two versions: the first one contains information about the nodes, such as the unique identifier (ID), the labels, the “x” and “y” coordinates that locate their position on the map, and the cluster number assigned to each node, among others. The second file corresponds to the network of relationships between nodes, which is essential for the creation of overlay maps. Subsequently, the 32,935 descriptors from the node file were loaded into an Excel spreadsheet, which facilitated sorting, relating, adjusting, and cleaning up equivalent terms, such as plurals, singulars, acronyms, acronyms, and inverted or similar compound words. For each descriptor, the term with the highest co-occurrence was selected to ensure consistency of the thesaurus. Periodically, the results of this cleaning were tested in parallel using VOSviewer to avoid line inconsistencies or errors. Once a preliminary base was obtained in Excel, the cleaning software OpenRefine and its text-faceting tool were used. The latter employs the “Footprint and Footprint n-gram” clustering algorithm, which allows similar terms to be displayed and corrected simultaneously against a preliminary Excel file. As a result of this process, the initial database was refined from 32,935 to 4,488 terms, forming the first preliminary thesaurus and leaving 28,447 remaining descriptors. Subsequently, the second cleaning stage generated a second thesaurus with 2,719 additional descriptors. Together, the two thesauri consolidated a file of 7,473 unique descriptors, while the final descriptor file contained 25,728 terms.