© Improve 2021. All rights reserved.
Big data and forensics
In their enormous amount of data, financial institutions continuously investigate suspicious financial transactions that (can) demonstrate criminal behavior and persons involved in it.
Within a financial organization in the Netherlands, for this so-called Forensic Economic Crime, customers are periodically screened against various lists made available by the company Accuity. These are lists with information about people and organizations that can indicate financial risk (politically exposed person, enhanced due diligence, sanctions). For this research RiskShield from Inform is in use. A web-based interface has been developed for analyzing the data and resulting alerts. Development takes place in three areas:
- Within RiskShield, Rules are developed, which are designed and implemented from an IT perspective, iteratively in small waterfall projects. These projects are aimed at reading information from various channels, such as the Accuity lists, but also customer data from the local systems. This information is stored in so-called matrices within RiskShield.
- Scrum teams, in combination with devops, develop the web-based interface in sprints that are as synchronized as possible with developments within RiskShield. Within RiskShield, the business logic is also developed, which is shielded from IT developments. This confidential logic is part of the entire process, but has a different approach.
- In an ongoing process, adjustments are made in small cycles based on signals from the market, ad-hoc requests from the business and progressive insight from the developers.
As a result of this combination of sub-areas, three different test methods have been chosen:
- Testing of predetermined requirements within the waterfall approach;
- Testing from the user stories within the scrum teams;
- A combination of exploratory testing and business impact analysis.
The complexity of the matter and unpredictability of transactions in the real production world, does not always make it possible for testing to use predictable behavior and test data. In addition to exploratory tests, the cycles also use business impact analyzes on the production flows. In addition to the functional criteria in the exploratory test, the workability of the implementation is also examined. How does the quality relate to the quantity of information obtained to be analyzed? By fine-tuning within the business impact analysis, a balance can be found here, in which the quality of the service for the customer comes first.