Visual exploration of cash flow chains

The paper proposes a new method for interactive visual exploration of the chains of financial transactions, assisting an analyst in the detection of money laundering operations. The method mainly concerns searching, displaying and annotating selected groups of transactions from a database. We show how one can programmatically and interactively reduce the volume of the chains surveyed and limit the analysis to the most suspicious transactions. In order to improve readability of the transaction graph, an evolution-based algorithm has been designed to optimize its visual representation. The system is verified on the real-life database of financial transactions. The experiments conducted have shown that allowing visual exploration, one can accelerate the search process and enrich the data analysis.