Performance analysis and a recommendation system in online analytical processing systems

Due to the ease and safety of digital media for data recordings, the usage rate is increasing day by day. The storage of data such as products, customers, location, sales time is especially important for the future decisions of the organizations. Variety of products and customer information make it difficult to process data. Therefore, an organization needs a system to manage, analyze and make large amounts of data available in decision-making processes. In this case, data warehouses can be used with analytical databases and on-line analytical processing (OnLine Analytical Processing, OLAP) tools to facilitate organizations in decision making process. Therefore, OLAP is a decision mechanism that many companies use and support in data-based systems. The increasing amount of data and the diversity of data every day have led to the need for more OLAP systems. Within the scope of this study, two different studies were conducted on OLAP systems: Performance analysis of OLAP approaches and design of a recommendation system in OLAP systems. Analyzes were made on a sample data warehouse about the performance of different OLAP engines, the duration of the query and in what situations they would be more efficient. As a result of these comparisons, MOLAP has the best performance in complex calculations, ROLAP is especially important in processing large amounts of data. The second study is to apply the recommendation systems on the OLAP, which prepare suggestions through various methods in many areas, such as films, music, news, books, search results, e-commerce, and the actions of the users on previous actions. Recommendation systems to be used in data warehousing systems, which are fed from historical data of organizations and are widely used in data science, are needed by managers and analysts. Designed with this study, the system offers a list of suggestions to the users through the combination of binary, collaborative filtering and categorical element algorithms. The system has been tested on the AdventureWorksDW data warehouse, which is available to everyone by Microsoft.