Literature review related to the data exploration methods in marketing showed that the number of publications and software tools in the field of exploration data analysis of business processes has been increasing since 2015. The contribution of the given research lies in the hypothesis of the possibility to construct a quantitative model of a business process for big data containing fraud. At the first stage, we performed data cleaning and reduction, examined data distributions, and identified issues such as extreme values that supplement the previously developed visualization approach. The descriptive analysis helped us to describe the significant number of categorical features. We emphasized the data outliers that should be marked for further research.