A Proposed Big Data Architecture Using Data Lakes for Education Systems

Nowadays, educational data can be defined through the 3Vs of Big Data: volume, variety and velocity. Data sources produce massive and complex data, which makes knowledge extraction with traditional tools difficult for educational organizations. Indeed, the actual architecture of data warehouses do not possess the capability of storing and managing this huge amount of varied data. The same goes for analytical processes; which no longer satisfy business analysts; in terms of data availability and speed of execution of queries. These constraints have implied an evolution towards more modern architectures, integrating Big Data solutions capable of promoting smart learning to students. In this context, the present paper proposes a new big data architecture for education systems covering multiple data sources. Using this architecture, data is organized through a set of layers, starting with the management of the different data sources to their final consumption. The proposal approach includes data lake as a means of modernizing decision-making processes, in particular data warehouses and OLAP methods. It will be used as a means for data consolidation for the integration of heterogeneous data sources.