The column-oriented data storage Vertica is used to manage large and fast-growing volumes of data in query intensive systems, e.g. data warehouses and other systems for data analytics. The speed of evaluating analytic queries largely depends on data projections (aka materialized views) defined. The student will study the area of column-oriented databases with a special focus on Vertica. The student will describe its query monitoring and projection tuning principles. The aims of the thesis is: (i) to analyze the Vertica Database Designer and its influence on different objects, (ii) to design and implement a testing tool targeting suboptimal designs generated by the designer, and (iii) to evaluate this tool on a TPC benchmark. This diploma thesis is done in cooperation with GoodData, s.r.o.