The scarcity and the heterogeneity of publicly available e-learning datasets are some of the main concerns for Learning Analytics research. For this reason, a common framework is needed to provide researchers with a richer information and a general schema, in which new datasets can be collected. The educational context can vary depending on educational level, national policies and cultural characteristics, so the integration of datasets is challenging and requires a dynamic approach. The thesis work focuses on the design of UNIFORM, an integrated open relational database for education, which guarantees modularity and flexibility: its structure can be dynamically edited, to allow the integration of new datasets.