Using OLAP Tools for e-HRM demography and absenteeism. This text represents the point of view of the IT staff regarding the project progress and sequels, intending to provide an example of how such powerful technology can help the Human Resource area, and potential dif- ficulties in this way. The projects began in 2005 at the Human Resources department of a public Brazilian University. It offers 58 undergraduate and 127 graduate courses and is organized in 20 Institutes and Schools, one academic medical center, 23 research centers and an administrative area. The institution employs teachers and researchers, technical and bureaucratic staff, physicians and nurses, comprising about 10,000 workers. This paper is organized as follows: in the next sections there are, respectively, an overview of the adopted Business Intelligence theory, the de- scription of implementations of two Data Marts regarding Human Resource information and how their content can be provided to final users. In the remaining sections, we discuss the challenges faced during the project, the drawbacks, lessons learned, project follow-up and present our final remarks. EXTRACT, TRANSFORM, AND LOAD The process of obtaining and modifying the data for feeding a Business Intelligence database is called ETL, in respect to the three steps it involves: Extract, Transform and Load. In the Extract step data are typically queried from other systems of the company, the so-called OLTP ­ On Line Transactional Processing ­ that supports the day-by-day organization operations. Spreadsheets and plain text files can also be used as data sources for Extraction. In the Transform step the data are handled aiming to fit in the view the users of the decision support system have of the process and of the facts they represent. This means unit conversions, codes standardization, data filtering, categoriza- tion and so forth. In the Load step the data produced by the prior steps are stored in a special database structure called Data Warehouse, which is described as follows. Business Intelligence Overview The concept of Business Intelligence (BI) refers to the abilities of the corporations to retrieve in- formation related to their operation processes and area of activity, in a flexible and dynamic way, allowing the analysis, detailing and understand- ing their work and providing means for decision support. The term has been popularized since the late 1980's by Howard Dresner and the Gartner Group (Power, 2002). The data managed by Business Intelligence systems have certain specific characteristics, reflecting on the way they are gathered, stored and retrieved, which will be briefly explained in the following sessions. DATA WAREHOUSE AND DATA MART The Data Warehouse (DW) is a large data reposi- tory (Inmon, 2005), obtained from all the relevant sections of the organization. The Data Warehouse contains the raw material for the management's decision support system. When the Data Warehouse is updated from ETL, no data is deleted or overwritten. Instead, the data are accumulated, constructing the history of the data involved in the company operations. The data structure of a DW often does not fol- low the common database systems techniques that use normalization to ensure data integrity and less storage space. Instead, the data are de-normalized and arranged in such a way that helps to query for reports and analysis. 253