Big data and analytics usage for anti-money laundering

Big data and analytics usage for anti-money laundering

Big data and analytics usage for anti-money laundering

Financial Intelligence Units (FIUs) are one of the biggest fighters against money laundering which is one of the biggest crimes worldwide, its amount equates as much as 5% of the global GDP1.  FIUs are key players between Law enforcement agencies and private sector to stop usage of illegal money and stop crime by usage of financial flow data which comes mostly from Suspicious transaction reports and Suspicious activity reports, but also from available information from other sources like external databases, business registers or simply open-source information.

Problem is that criminals become more and more intelligent as well by using cross border payments, cryptocurrencies, dark net or other anonymization services so that it becomes extremely difficult to track. Offen number of transactions are so big with so many threat actors that analyst have to rely on various disparate identifiers or it can concentrate on small interesting data point, but lose his possibility to find bigger picture, just because they don’t have appropriate tools and infrastructure.

Most of FIUs have analytical tools for data analysis, but background situation changes very quickly so these tools require regular feature improvably and also modular improvement so that intelligence units can connect information from internal data storages and external sources and also put intelligence on top of these data. That mean that it is not enough to extract data from various sources and load the in one database, it must be common data warehouse or even data lake which allow regular data updating and change monitoring so that it is possible to detect suspicious transactions or possible illicit actors with Machine learning methods, so that during this analytical process it is possible to improve data analytics. Reacher’s already seeding for best mechanisms to build predictive monitoring system for anti-money laundering2.

FIU Latvia is also seeking for best ways to analyse and aggregate several data sources. For better performance and possible data reusage in long term period we are trying to build distributed computing environment with business intelligence tools, so that full intelligence cycle can be provided between multiple tools as one common mechanism. Basically, main objective is to build system from multiple layers so that each layer helps to get extra value from extracted data. Main layers are – storage and computational level, data integration level, analytical layer and user interface layer.

FIU Latvia understands that main partners for fighting crime are Law enforcement agencies, so that is very important to understand their requirements as well, so that such system can bring extra value also for them. As great platform for cooperation with intelligence practitioners and security specialists from several law enforcement agencies and academic institutions is NATIONES project ( This project allows monitor about newest researches and achievements applied in security and intelligence, develop common needs for future improvements by connecting practitioners and academics, industrial bodies and professionals.  FIU Latvia participates in working group which has started in February 2022 and these groups come together to discuss digital needs and problems which should be solved with modern technology usage our which barriers should be solved so that data exchange and analytical mechanisms could work much better and faster. In this working group we also discuss possible solution for common data analytical platform which could be used for intelligence processes in law enforcement agencies.


Author: Nauris Paulins, Deputy head of Innovation and IT division of Financial Intelligence Unit of Latvia


Office on Drugs and Crime. Money Laundering.

Tertycnyi, M. Godgildieava, M. Dumas, M Ollikainen. Time-aware and interpretable predictive monitoring system for Anti-Money Laundering, 2022. Machine Learning with Applications, 1-11, Article 100306.