Link Analysis

7 min read

Link analysis is a way to make use of very large datasets from different sources. It does so by visualizing the relationships between different data points, such as people, organizations, events, or transactions. It shows these connections as nodes and links, and helps you find insights that are not obvious in traditional data analysis, revealing the hidden connections and relationships that matter. For example, link analysis can help spot fraud, detect anomalies, discover new trends, or understand complex networks. Link analysis is useful for many fields, such as intelligence, law enforcement, security, and journalism, where you need to know the context and meaning of your data. Link analysis can help you see your data in a new way, and reveal the hidden secrets that matter.

By using link analysis, you can see your data as a network, where each node is a data point, and each link is a connection. Link analysis can help you discover the principles that govern your data, such as:

  • How are the nodes and links distributed? Are there clusters, hubs, bridges, or outliers?
  • How are the nodes and links related? Are there similarities, differences, dependencies, or influences?
  • How are the nodes and links changing over time? Are there patterns, trends, cycles, or anomalies?

Link analysis has three primary purposes:

  • Discover connections that match existing patterns of interest among linked entities.
  • Detect deviations from expected patterns that indicate anomalies.
  • Identify new and interesting patterns that emerge from the data (for example, in social media and business domains).

If you have a lot of data, you have a lot of potential knowledge. But there might be hidden insights and connections that you are still missing, especially when using table-based databases. Table or document-based systems cannot answer complex questions, especially in the presence of unknowns. Multi-hop queries are slow, never return results, or crash the system. Graph-at-core link analysis has many advantages:

  • Thorough data collection: from various structured and unstructured sources. Link analysis lets you connect information from different sources, for a more complete view of information.

  • Consistent data cleansing and entity extraction: this involves making the information consistent and unique, and ensuring nodes and links have the right attributes and identifiers for your analysis.

  • Custom data model construction: this is the process of turning a conceptual model, based on the information you want to find out, into a logical model, based on the information you have.

  • Intuitive visualization and analysis: this is when you view and interact with nodes and links in a link analysis chart. You can enhance your visual analysis with various techniques and algorithms, such as centrality analysis, automated layouts, timeline visualization and node grouping.

AML (Anti-Money Laundering) is an important field in modern criminal investigation. Money laundering enables criminals to evade taxes, fund terrorism, and finance other illicit activities. Link analysis is a powerful tool for anti-money laundering (AML) investigations, as it can help you follow the money, identify the actors, and understand the context of the transactions.

  • How are the nodes and links distributed? Are there clusters, hubs, bridges, or outliers that indicate suspicious activity or risk?
  • How are the nodes and links related? Are there similarities, differences, dependencies, or influences that reveal patterns or anomalies?
  • How are the nodes and links changing over time? Are there trends, cycles, or deviations that signal changes in behavior or risk?

Big data is everywhere. It’s the massive amount of information that is generated, collected, and stored by various sources, such as social media, sensors, transactions, and more. But big data is not just about quantity. It’s also about quality. And that quality depends on how well you can analyze and understand the data. Link analysis can help:

  • Find the relevant information in a sea of data. Link analysis can help you filter, highlight, or expand the nodes and links that are of interest to you, and hide the ones that are not. This way, you can focus on the data that matters, and avoid getting overwhelmed by the data that doesn’t.
  • Understand the complex relationships within the data. Link analysis can help you see the structure and meaning of your data by looking at the nodes and links, and their properties. You can also explore the data interactively, by clicking, dragging, or zooming on the nodes and links, to discover new connections and patterns that are not obvious in other forms of data analysis.
  • Extract meaningful insights from the data. Link analysis can help you identify trends, anomalies, or associations in your data, by using your visual link analysis, rather than relying on complex algorithms or calculations. You can also use link analysis to compare and contrast different data sources, to gain a holistic view of your data.

Link Analysis for Big Data

Link analysis algorithms are methods that use mathematical and computational techniques to analyze connections between entities in a network. They can help us measure the importance, similarity, or influence of nodes and links, and find patterns or anomalies within the network. Link analysis algorithms can be applied to various domains and scenarios, such as web search, fraud detection, and criminal intelligence. Link analysis software, such as GraphAware Hume, goes beyond link analysis and provides a comprehensive platform that integrates data from various sources, supports search, entity extraction, querying, and collaboration, and leverages graph database to deliver faster and better results.

By using link analysis software, intelligence analysts can visualize and explore networks of entities and relationships, such as suspects, locations, and transactions. It can help them to solve various types of cases, from cybercrime, to human trafficking. For example, link analysis can help identify the structure and hierarchy of criminal organizations, track the flow of money and drugs, and discover new leads and evidence. Link analysis can also provide real-time situational awareness and risk assessment, which are crucial for the safety and effectiveness of law enforcement operations.

To combat fraud, investigators need to find and analyze hidden connections and patterns among the actors and transactions involved in fraudulent schemes. By using link analysis fraud detection software, such as GraphAware Hume, investigators can detect various types of fraud, such as insurance fraud or money laundering, by finding matches, outliers and other indicators of suspicious behavior. Link analysis can also help verify the information and sources used in fraud investigations, and provide evidence and insights for prosecution and prevention.

  • Law enforcement: Link analysis can help law enforcement agencies to find and monitor criminal networks, such as organized crime, drug trafficking, human trafficking, and terrorism. By using link analysis, they can see the structure and hierarchy of criminal groups, locate the sources and targets of criminal acts, and find new clues and evidence.

  • Anti-fraud: Link analysis can help detect and stop various types of fraud, such as insurance fraud, money laundering, identity theft, and tax evasion. By using link analysis, they can identify anomalies and suspicious patterns in the data, such as matches, clusters, and outliers. They can also check the information and sources used in fraud investigations, and provide evidence and insights for prosecution and prevention.

  • Security and intelligence: Link analysis can help security and intelligence agencies to discover threats and vital intelligence, such as cyberattacks, espionage, and sabotage. By using link analysis, they can study the connections and behaviors of enemies, from hackers, to terrorists. They can also provide real-time situational awareness and risk assessment, which are essential for the safety and effectiveness of security operations.

  • Business analytics: Link analysis can help businesses to obtain insights and improve their performance, such as marketing, sales, customer service, and innovation. By using link analysis, they can learn the preferences and behaviors of their customers, such as their purchase history, loyalty, and satisfaction. They can also recognize the influencers and trends in their market, such as their competitors, partners, and opportunities.

Graph analytics and link analysis are two related techniques that involve the visualization of connected data. However, they are not exactly the same. Graph analytics is a more general term that covers not only the visualization, but also the analysis and manipulation of graph data, such as finding shortest paths, clustering, and more. Link analysis, on the other hand, is more specific and concentrates on the visualization and interpretation of the connections and relationships within the data, such as identifying patterns, anomalies, and trends. In other words, link analysis is a part of graph analytics, and both use the same graph data model.