A Business Intelligence Dashboard Design Approach to Improve Data Analytics and Decision Making

Abstract

This paper considers a problem of dashboard design in order to improve data analytics and decision making as business intelligence capabilities. Essentials of data warehouses and data marts used as data sources for business intelligence activities are briefly outlined, and the star schema data structure is considered as the most widely used. Data visualization is considered from the perspective of proper visualization graphs and charts selection. It is extremely vital for designed dashboards, since inappropriate visuals may mislead users and shift their focus to wrong things. Bar charts, line charts, and pie charts are considered as the most common visualization graphs. Proposed approach includes two phases: dataset preparation and dataset analysis. While dataset preparation is mostly focused on star schema transformation into flat structures, dataset analysis phase proposes recommendations on which visualizations may be placed on a designed dashboard. In order to propose such recommendations, threshold values of dataset sizes are used. A dashboard design process, which is considered as the baseline of the proposed dashboard design approach, is outlined. Sample dataset in considered, five data subsets are prepared and recommendations on visualization charts for these datasets, which may be placed on a dashboard, are proposed. Obtained results are discussed, conclusions are made, as well as the further research objectives in this field are formulated.