A systematic approach for Systematic Dataset Reviews

Abstract

This research introduces guidelines for performing a Systematic Dataset Review (SDR), a term newly defined within this research. The SDR guidelines tackle the issues currently faced with dataset research. Currently, documentation is often inadequate or lacking in research that uses datasets, especially regarding the steps taken in finding and handling datasets. The SDR guidelines aim to tackle this issue by providing researchers with guidance during their study, allowing them to perform and document systemically. The SDR guidelines are developed using various research methods. First, existing literature was studied to identify the existing knowledge regarding research involving datasets. Then, interviews were held with experts in the field of information and computer sciences, that have experience with using datasets within their research. Afterward, the created SDR guidelines were evaluated by doing an expert interview and by using the SDR Guidelines within example research. The results of this research are the guidelines for performing a Systematic Dataset Review. The guidelines consist of 2 phases, each with distinct activities. The first phase is the planning phase, which not only helps guide researchers in planning their research but also helps document these steps in a concise systematic way. The second phase is the conducting phase, which encompasses performing the corresponding steps from the planning phase and altering them if necessary. The SDR Guidelines are most suited for novice researchers, such as Bachelor, Master, and PhD Students who do research that requires the use of at least 1 dataset.