In this paper, we focus on the proposition of new computational mechanisms for the individual and subsequent steps detection in the financial flows analysis. In other words, for selected starting points (e.g. bank account, date, value, etc.), the analysis aimed to identify further actions taken by the perpetrators. Such solution is much needed for the police, forensic experts and analysts of financial crimes and transactions, who need to manually analyze big complex data. In this paper, we show the application of our proposed method and solution for two financial Internet-related crimes, namely financial pyramid and on-line shopping fraud.