Search

Advances in technology have enabled organizations to collect a variety ofdata at high speed and provided the capacity to store them. As a result theamount of data available is increasing daily at a rapid rate. The data stored inorganizations hold important information to improve decision making andgain competitive advantage. To extract useful information from these hugeamounts of data, special techniques such as data mining are required. Datamining is a technique capable of extracting useful knowledge from vastamounts of data. The successful application of data mining in organizationsdepends on several factors that may vary in relation to the environment. InMozambique, these factors have never been studied. The study of the factorsaffecting the use of data mining is important to determine which aspectsrequire special attention for the success of the application of data mining.This thesis presents a study of the level of awareness and use of datamining in Mozambique and the factors affecting its use. It is a step towardsthe development of a framework to facilitate the application of data miningin Mozambique. The study is exploratory and uses multiple case studies intwo institutions in Maputo city, the capital of Mozambique, one in the areaof agriculture and the other in the field of electricity, and of Maputo citymore broadly. The study involved a combination of observations, focusgroup discussions and enquiries directed at managers and practitioners onaspects of information technology (IT) and data analysis. The results of the study reveal that the level of awareness and use of datamining in Mozambique is still very weak. Only a limited number ofprofessionals in IT are aware of the concept or its uses. The main factorsaffecting the use of data mining in Mozambique are: the quality, availabilityand integration of, access to data, skill in data mining, functional integration,alignment of IT and business, interdisciplinary learning, existence ofchampions, commitment of top management, existence of changemanagement, privacy, cost and the availability of technology. Threeapplications were developed in two real settings, which showed that thereare problems to be solved with data mining. The two examples in the area ofelectricity demonstrate how data mining is used to develop models toforecast electricity consumption and how they can enhance the estimation ofelectricity to be sold to the international market. The application in the areaof agriculture extracts associations between the characteristics of smallfarmers and the yield of maize from a socioeconomic database with hundreds of attributes. The applications provide practical examples of howdata mining can help to discover patterns that can lead to the development ofmore accurate models and find interesting associations between variables inthe dataset. The factors identified in this thesis can be used to determine thefeasibility of the implementation of data mining projects and ensure itssuccess.