Business Intelligence (BI) is the process of transforming your data into usable and useful information. This process is quite complex due to the diversity of informations sources available. Up to now the common approach to setting up a business intelligence process consisted in a fixed step process consisting in (1) defining a target data model, (2) setting up a data-warehouse implementing the model, (3) extracting and cleaning data from a multitude of data sources, (4) populating the data-warehouse using the, (5) building reports with the aggregated data. Multitude of software solutions exist for each of these steps.
However, relying on a data-warehouse mostly implies relying on a relational database. This type of technology is well adapted when it come to very well defined data and copes very well with data which is very well structured. Unfortunately, much precious data does not easily fit into a relational database since it is not sufficiently structured. In a database oriented BI process this either means excluding such data from the process, sacrificing the unstructured parts of the data or putting an enormous effort in cleaning the data.
Among others, on of the objectives of the ObjectML project is a effort to create Business Intelligence solutions allowing to cope with multiple forms of data. It is an object-wharehouse allowing to store business objects with different levels of structuredness : from very structured data to poorly structured data. It allows to efficient searching of the data by combining information retrieval techniques with classical boolean logic as implemented in database software.