Using graph data analysis to combat financial crime

Forensic analysts frequently deal with highly interconnected data stored across multiple-siloed IT systems. When the data is held in traditional relational databases, the relationships between data points are not apparent and need to be reconstructed to obtain meaningful insights. When a problem requires analysis of large volumes of connections with multiple levels of separation, solving the problem can be a complex task.

When performing analysis using relational databases, it is possible to connect data points to each other one at a time. In comparison, with graph database analysts are able to make multiple connections simultaneously and identify a network of connections.

Graph technology uses algorithms and queries to analyse a graph database of structured and unstructured data to connect multiple data points and identify relationships between them that might not otherwise be discovered, or which would take longer to discover, through analysis of relational databases. Graph analysis provides a holistic view of entities relevant to an investigation – such as people, companies and transactions. This helps investigators to uncover even complex criminal schemes and networks.

Both internal and external data, and structured and unstructured data, can be accessed through graph analytics software, relating to payments and other transactions, customer data (including email, SSN, IP) and third party data (blacklists, social media, company registries etc.) and employee data. The ability of software to analyse unstructured data is evolving rapidly.
The output from an investigation is presented in a graph form, highlighting connections between data points that have been discovered. Visual presentation, which can be customised to the user’s requirements, helps to clarify the connections that have been discovered.

Graph analytics enables investigators and financial crime compliance teams to carry out analysis more efficiently as well as more effectively. A graph solution of connected data enables users to run complex network analytics searches. This can be particularly valuable in saving time at the early stage of an investigation when all the information (or accurate information) is not already on hand.

Graph analysis reduces the time investigators spend on low-value tasks, such as manually reconstructing connections between entities within tabular data or by gathering data from different systems. Having identified connections more quickly, this can leave more time for carrying out investigations themselves.

The use of graph technology can also speed up significantly the investigation and discovery by financial institutions of financial crimes. For example, one banking client of Deloitte noted that using the technology had accelerated the task of exploring connections. In some cases, the client was able to identify complex hidden connections between multiple entities within just a few seconds. The same analysis would have taken days to perform with traditional tools based on tabular information.

Interfaces for graph analysis are designed for use by non-technical as well as technical users, providing user-friendly access to the data graph’s capabilities. Interfaces also allow for a high level of customisation, enabling users to adapt graph visualisations of multiple connections to their particular needs, without the need for technical expertise.