Article Preview
TopSome authors (Bimonte, Sautot, Journaux, & Faivre, 2017) propose strategies to make easier the process of designing and building a data warehouse (Chandra & Gupta, 2018; Ralph. Kimball & Ross, 2013; Romero & Abelló, 2009), Others suggest ways to keep the track of the whole history of objects in data warehouse efficiently (Atay & Garani, 2019; Golfarelli & Rizzi, 2009). However, reaching the bond among data from different thematic contexts, existing in these repositories, is still a challenge.
Data marts, which compose data warehouses, are non-volatile data repositories, oriented by subject or themes (R Kimball, 2012; Ralph. Kimball & Ross, 2013), being shaped according to Star schemas or Snowflake schemas. These schemas were created to facilitate the manipulation and visualization of large volumes of data. Star schemas follow a denormalized data approach whereas snowflake schemas follow some rules of normalization (Garani & Helmer, 2012).
This work applies the dimensional schema, used in data marts, to find relations (associations, correlations and causalities) in the stored data. It was not found in the areas of Business Intelligence, Data Analytics or Data Mining a strategy that uses the structure of star schemas to find and evaluate relations in an automatic and comprehensive way, trying to find relevant connections, as proposed in this work.