An Open Dataset Architecture Prototype to Drive Innovative Data-Driven Decision

Abstract This study developed, implemented, and validated a prototype open-dataset architecture to promote innovative data-driven decision-making (DDDM) in Uganda's tertiary education sector, with the Uganda Institute of Information and Communications Technology (UICT) as a case study. Despite increased investments in digital transformation initiatives and educational information systems, institutional challenges such as fragmented data systems, limited interoperability, weak analytics integration, and underutilization of institutional datasets remain unresolved. Following the Design Science Research (DSR) methodology, the study employed a mixed-methods approach, including assessments of institutional requirements, platform prototyping, implementation, and validation. Data were gathered through baseline and post-implementation surveys, stakeholder interviews, document analysis, User Acceptance Testing (UAT), and validation workshops. Quantitative analysis used descriptive statistics and Pearson correlation analysis, whilst qualitative responses were examined thematically. The findings found that UICT's data ecosystem was extremely fragmented among systems such as ACMIS, LMS, HRMIS, and finance platforms, resulting in inconsistent accessibility, limited data integration, and event-driven rather than continuous data utilization for institutional decision-making. The prototype had metadata-driven dataset management, role-based access controls, analytics dashboards, advanced search and filtering tools, export capabilities, and a secure, modular architecture coupled with Power BI and Python Dash analytics tools. Usability testing revealed good platform performance, with a System Usability Scale (SUS) score of 78.4, exceeding the industry norm of 68. This was accompanied by high job completion rates and positive stakeholder views of accessibility, analytics capabilities, and workflow alignment. Although statistical testing revealed no significant relationship between perceived platform impact and data-driven decision-making frequency during the pilot period, qualitative findings indicated significant practical benefits, including improvements in institutional analytics capacity, strategic planning, evidence-based governance, and innovation. The study indicates that the proposed architecture offers a scalable, context-responsive framework to improve digital transformation and data-driven governance in Ugandan higher education institutions and other low-resource settings. Keywords: data-driven decision-making, open dataset architecture, digital transformation, higher education analytics, institutional data governance.