No-code entity resolution and graph technology: The key to next-gen investigative analytics

Analyzing relationships in data is critical across countless applications, from detecting accounts linked to money mules, to tracking down networks of fraudsters, to identifying cybersecurity threats. Investigators and analysts have tools at their disposal to manage their data, but all too often those tools fall short when it comes to analyzing complex networks of relationships. Those shortfalls become even more apparent when dealing with multiple systems and sources.

Graph-based technology can overcome the limitations of traditional tools, building a single source of truth of connected data. But here lies another challenge: merging multiple data sources generally results in messy data sets full of duplicates and missed links.

That’s where entity resolution comes in. Entity resolution helps build a unified knowledge graph that lets investigators sift through complex data and surface key insights faster and more efficiently. Many entity resolution techniques are tedious and expensive - but what if there was an integrated, user-friendly solution?

We met with Paco Nathan, Dev Rel at Senzing, to talk about the importance of entity resolution, how it complements graph analytics and visualization for investigative use cases - and how the no-code solution available in Linkurious is making it more accessible.

Entity resolution - also called data matching - is the process of identifying and linking different representations of the same real-world entity across multiple data sources. Consider this common scenario: records for a customer across two different databases may show slight variations in identifying information - Bob R. Smith II vs Robert Smith Jr., for instance - but actually refer to the same person.

In a world where organizations rely on massive amounts of data from diverse sources, entity resolution is becoming indispensable. 

“There is so much demand to be using data, to have data driven processes,” said Paco Nathan when we interviewed him during Linkurious Days London. “But you need to have your ducks in a row. You need to have data that's usable. You need to be concerned about provenance and quality.” Entity resolution helps ensure data is high-quality and reliable, creating a foundation for accurate analysis and decision-making.

Without proper entity resolution, organizations face fragmented data, where the same entities appear as separate records, leading to incomplete analysis and missed connections that could be crucial for investigations.

The relationship between entity resolution and graph technology is naturally symbiotic. “One of the things we found is that wherever there are entity resolution practices, generally they're adjacent to graph practices. And certainly we find that there's this assumption in entity resolution that your data is somehow connected. So it's natural that it would fit into a graph,” explains Paco Nathan.

This connection makes perfect sense when you consider that entity resolution is fundamentally about understanding relationships - identifying when different data points refer to the same entity. Graph technology excels at representing and analyzing these relationships, making it a natural solution for working with entity-resolved data. “By running entity resolution, the artifacts that you get out of it are actually graph elements,” says Paco Nathan.

Graph analytics and visualization tools provide the framework to explore the complex networks of connections between resolved entities. When entities are properly resolved, the resulting graph becomes a more accurate representation of reality, enabling more sophisticated analysis and pattern detection.

Entity-resolved knowledge graphs serve as a single source of truth for investigators working on complex cases. Consider the kinds of investigations that benefit from this approach, to give just a few examples: 

• Networks of money mules moving funds across multiple accounts
• Cyber fraud schemes involving numerous fake identities
• Insurance fraud rings where the same individuals appear under different names across various claims
• Corporate investigations tracking beneficial ownership through layers of shell companies.

In these scenarios, an entity resolved knowledge graph acts as a comprehensive foundation where hidden relationships become easier to track down thanks to increased data quality and reliability. Graph analytics and visualization offer a natural way to connect the dots, shining a light on the relationships within complex data that might otherwise remain buried in separate systems.

Using graph technology to act on entity resolved data, the result is more accurate analytics, alerting, and investigative work. Investigators can work more efficiently, following connections that would be impossible to detect in fragmented, unresolved datasets. Patterns emerge more clearly, suspicious activities become more apparent, and the time-to-insight is dramatically reduced.

The importance of entity resolution shines through in financial crime investigations. Financial institutions rely on both internal and external data to combat financial crime like money laundering and fraud. When investigating suspicious activity, it's vital to know, for example, whether there are three people performing one transaction each, or rather one person performing three transactions - a distinction that could mean the difference between routine activity and suspicious behavior.

These organizations work with multiple data sources: they have their internal databases, of course, but may also rely on external sources like company registers, sanctions lists, and OSINT (Open Source Intelligence) data. Entity resolution can connect data relating to the same entities across these different databases, eliminating duplicates and providing investigators with a complete picture of potentially suspicious activities and relationships.

There are many ways to perform entity resolution, but not all approaches are created equal. Simple solutions are easy enough to build but lack accuracy when the data starts to get complex. Building your own comprehensive solution can quickly become costly, both in terms of development time and ongoing maintenance.

The consequences of inadequate entity resolution are significant. 

“The reality is if you're not going through this step of doing entity resolution or you perform it poorly, you're going to end up with what I would call balkanization in your data. Things that should be connected will not be connected. People who have an obvious association are separate. And when you try to apply graph algorithms, like centrality analysis, if you don't have those connections made in the data beforehand, the algorithms perform poorly,” warns Paco Nathan.

This “balkanization” creates a fragmented view of reality where related entities appear as isolated nodes in your graph. Critical connections remain hidden, making it impossible to identify the very patterns and relationships that investigators need to uncover. The sophisticated graph algorithms that could reveal important insights—such as centrality analysis to identify key players in a network—simply cannot function effectively without properly resolved entities.

Recognizing the critical importance of entity resolution for effective graph investigations, Linkurious has partnered with Senzing, a provider of cutting-edge entity resolution technology, to bring enterprise-grade entity resolution directly to the Linkurious platform. 

The solution is now available via a no-code interface, ensuring not only accuracy but also ease of use. This approach makes entity resolution accessible to all users, regardless of their technical background. Investigators and analysts can now perform sophisticated entity resolution without needing to write code or manage complex technical implementations.

The result is that users can resolve their data directly within Linkurious and immediately investigate it as a graph. This integration ensures transparency and accuracy for better, faster graph investigations. And it makes advanced investigative capabilities accessible to end users who don't have technical expertise but have deep domain knowledge about their investigations.

By integrating Senzing's innovative entity resolution technology into Linkurious’s intuitive graph visualization and analytics platform, organizations can now unlock the full potential of their investigative data without the traditional barriers of complexity and technical requirements.

Watch the full interview with Paco Nathan below. You can also watch the episode of his Graph Power Hour podcast featuring Linkurious here.