In this case study we’ll take a closer look at Unit21, a Google-backed tech company who use ReGraph and KronoGraph to transform the way companies deal with financial crime detection and investigation.
Financial crime: a big data problem
Detecting financial crime is a complex challenge. Criminals change tactics and cover their tracks to stay under the radar. The analysts and investigators tasked with stopping the perpetrators use data – lots of it – to uncover patterns of behavior, spot anomalies and reveal suspicious activity.
It’s a tried and tested approach that requires specialized knowledge. Many banks and law enforcement agencies hire teams of data experts to work alongside fraud analysts. These experts wrangle data and write complex queries to uncover unusual data patterns. But not every organization has a department full of highly-trained data scientists and software engineers. And fraud analysts shouldn’t be expected to learn SQL or create complex modeling tools: they’re experts in fraud and analytical techniques for detecting crime.
So who fills the gap? Say hello to Unit21.
Unit21: accessible financial crime management
Founded in 2018, Unit21 gives risk, compliance and fraud teams a better way to fight financial crime. Their customizable risk and compliance management platform puts analysts and investigators in the driving seat. It gives them intuitive ways to engage with financial data and make efficient decisions – all without writing a single query or line of code. So far, they’ve helped their customers monitor over $100bn in transactions.
Recently, the Unit21 team added ‘graphical link analysis’ to their platform’s suite of tools. Powered by ReGraph and KronoGraph, these components provide link and timeline analysis capabilities that make complex financial crime data simpler, clearer and more intuitive than ever.
Let’s take a closer look.
Building a better financial crime management workflow
Unit21’s platform uses a familiar ‘flag and review’ workflow to surface suspicious activity.
Analysts load their data and select which pre-written, self-training, configurable detection models they want to run. These flag up unusual events and connections, which are assigned to tickets ready for analyst triage. It sounds simple, but the potential complexity is mind-boggling.
Customers can import data in any number of formats, either streamed or batched, containing whatever entities and properties they choose. Their use-cases vary too, from transaction monitoring and fraud detection to identity verification and regulatory compliance. No two customers or data models are the same.
That’s where the power of visualization comes in.
The Unit21 platform harnesses ReGraph-powered graph visualization to visualize entity relationships, and KronoGraph-powered visual timeline analytics for transactional data. The graph and timeline models are flexible enough to work with any data at any scale.
Nick, Software Engineer at Unit21, explains:
Visualizing entity relationships: spotting unusual connections
The first thing an analyst looks for in a new financial crime dataset is overlapping entities. Whether it’s a shared address, device or account, it’s important to see these unusual connections. It’s the perfect task for graph visualization.
Here’s a simple dataset showing people, accounts and money flows in Unit21’s network analysis chart, built with the ReGraph SDK:
Icons and colors help differentiate between node (entity) types, while solid and dashed links show the relationships between them. ReGraph’s layout algorithms ensure the network is neatly distributed across the chart, regardless of the size or shape of the dataset.
One of the most powerful ReGraph features harnessed by Unit21 is combos: the ability to group nodes and links by any attribute. Nick explains why:
Here, nodes represent bank accounts, grouped by status depending on whether they’re approved, active or on a watch list.
ReGraph also complements the platform’s powerful labeling approach. Customers can import near-limitless numbers of data attributes, then easily add them to charts as labels or use them as filters.
Here, for example, we’re filtering out nodes associated with a single IP address. It’s good practice to remove unhelpful or unnecessary nodes from the view. Filtering this way keeps the chart tidy and helps the analyst find their answers faster:
Visualizing transactions data: going beyond spreadsheets
Alongside their graph views, the Unit21 team uses sophisticated visual timelines to understand transactions and money flows. Traditionally, much of this kind of analysis happens with spreadsheets, or aggregated analysis which fails to show individual transactions:
For Unit21, again, flexibility was key. With customers ranging from traditional banks to cryptocurrency exchanges, they needed a visualization approach that would shed light on any transaction pattern in any currency.
KronoGraph’s simple yet flexible model was ideal. Data entities, like accounts, currencies and instruments, are listed on the left-hand side. In the timeline, transactions are the connections between entities:
KronoGraph’s grouping capability is used to good effect, too, showing the base account in one group and counterparties in another. This is especially useful for showing money flows around systems that eventually return to a single account – often a sign of fraud rings.
Along the top of the timeline, a set of filters lets the user add or remove data from the view. Advanced grouping options nest data attributes, like status and subtype (e.g. currency) beneath entities. The result is a clear and simple user experience that empowers analysts to explore complex transaction data at their own pace and scale – easily, intuitively, and without a single line of code.
This powerful visual analysis functionality is steadily being rolled out to every Unit21 customer. Meanwhile, their engineering team has already started to develop new and more advanced visualizations, helping them bring customizable, no-code financial crime management tools to a wider audience.
In ReGraph and KronoGraph, we found visual data analysis capabilities that help our customers get the most out of their data, and super-charge their detection and investigation workflows. The flexibility, scalability and performance of the toolkits makes them by far the best solutions on the market. We’re excited about using them to create even more innovative and insightful visual approaches to data exploration and analysis.
For more information about Unit21, visit www.unit21.ai