EDMLENS DATAFLOW: AN ANALYTICAL ARCHITECTURE FOR EXTRACTION, TRANSFORMATION, LOADING, ANALYSIS AND VISUALIZATION OF EDUCATIONAL DATA

Over time, educational institutions have adopted new technologies or improved those they already used, thus generating large volumes of data. This data can help in the process of choosing certain strategies to improve teaching. However, there are many situations in which data is stored but is not used by management or other professionals to support decision-making. In some cases, this may occur due to the difficulty in finding a data analysis tool or method that does not require specific knowledge for its use. For this reason, this paper presents a bibliographic review of the literature with the objective of supporting the development of a data analysis architecture, called EdmLens DataFlow, which integrates processes of protection, transformation, loading, analysis and visualization of educational data. This architecture offers a holistic structure that, in addition to facilitating data interpretation, improves informed decision-making, promoting a monitoring and planning process. Therefore, the literature review covers the areas of Data Science, Knowledge Discovery in Databases (KDD), Data Analysis Architecture and Academic Analysis, and highlights the following elements as results for the development of EdmLens DataFlow: data protection from different sources and formats, working with structured, semi-structured and unstructured data; data transformation to make them all structured, performing normalization, cleaning and integration; loading the data into a Data Warehouse, as this storage approach works with structured data; data processing and analysis through artificial intelligence and machine learning; and data visualization, based on the analysis, using reports and interactive dashboards.