Spotlight on Graph Analytics in Finance

<p>Think of ‘graph analytics’ (GA) and you might think of quants poring over bar graphs and pie charts. In fact, GA is something else entirely: a powerful tool for mapping networks, relationships and interactions between people, organizations and objects. So far GA’s most celebrated uses are in analyzing social media data. But it is in the financial services industry that the technology could be transformational, especially when used to underpin Artificial Intelligence (AI)-based analytics: stabilizing Know Your Customer (KYC) processes, for example, or lessening time-consuming activities such as data processing, validation and error checking.</p>

<p>In this Spotlight, written in collaboration with our research partners Quantexa, we :</p>

<ul>
<li>Define GA, providing an overarching taxonomy.</li>
<li>Consider its strengths and weaknesses as an analytical technique.</li>
<li>Examine GA success stories in financial services.</li>
<li>Consider areas in which GA will be used in the future.</li>
<li>Explore how firms should establish and embed successful GA processes.</li>
</ul>

<p>GA represents an evolution in analytical techniques toward the analysis of networks.</p>

<ul>
<li>Events such as payment requests can be <strong>scored independently</strong> using an analytics system. But they may represent only a small part of the total available data, so the system could generate inaccurate results.</li>
<li>Pulling in <strong>data on the entities</strong> connected to the event – the individual or company concerned, or the address where the event occurred – can increase the accuracy of the data model. However, this can add significant complexity.</li>
<li>Going one step further, and <strong>mapping networks of relationships</strong> that exist between actors across related events – such as loan applications and insurance claims – can provide a higher-definition view of the incident, and greatly increase the quality of predictions. Such network visualization requires GA capabilities to operate efficiently.</li>
</ul>

<p>By using GA effectively, Financial Institutions (FIs) can gain invaluable – and rapid – insight into their systems (via cyber risk management), their counterparties (via counterparty credit risk), their customers (via KYC/Anti-Money Laundering [AML]), and the wider world (via supply chain analysis).</p>

<p>When considering GA, however, prospective users should consider the following factors:</p>

<ul>
<li>Unlike other analytics tools, GA enables FIs to analyze the connections and relationships between entities. However, GA is not a catch-all technology – only when it is <strong>applied to the correct use cases</strong> does it demonstrate its utility.</li>
<li><strong>The adoption of GA </strong>has been supported <strong>by recent innovation</strong> in areas such as High-Power Computing (HPC), the development of specific graph databases, the use of large-scale&nbsp;unstructured data stores, and the use of GA in combination with techniques such as entity resolution and Artificial Intelligence (AI).</li>
<li>Despite a poorly defined Return on Investment (ROI) argument for many next-generation analytics, <strong>GA is proving its value</strong>. Graphs are being used to measure the sanctions or fraud risk connected with individuals in commercial and retail banks, and in the areas of business development, credit risk and analyzing supply chain networks.</li>
<li><strong>Alternative (aka ‘alt’) data</strong> can also be used to inform investment decisions and hedge fund analysis.</li>
<li>If FIs use GA they should <strong>involve stakeholders</strong>, and ensure they have relevant domain expertise, as well as knowledge of the trade-offs associated with constructing graphs.</li>
</ul>

<p>GA has significant potential. By combining it with other analytical techniques, it’s possible to develop an AI capability to either automate decisions or augment and improve human decision-making. Recently, the focus on cutting-edge analytics has been on deep learning and Machine Learning (ML) techniques, but GA can be equally transformative in its effect on financial and risk technology.</p>

<p>GA and ML are also fundamentally complementary techniques, and there are many use cases to which they can be applied. In tandem, they can add significant value in the area of counterparty credit risk, expanding into the commodity and alt-data chain and cyber security analysis.</p>

  • LinkedIn  
  • Save this article
  • Print this page  

Only users who have a paid subscription or are part of a corporate subscription are able to print or copy content.

To access these options, along with all other subscription benefits, please contact info@risk.net or view our subscription options here: http://subscriptions.risk.net/subscribe

You are currently unable to copy this content. Please contact info@chartis-research.com to find out more.

Sorry, our subscription options are not loading right now

Please try again later. Get in touch with our customer services team if this issue persists.

New to Chartis Research? View our subscription options

You need to sign in to use this feature. If you don’t have a Chartis account, please register for an account.

Sign in
You are currently on corporate access.

To use this feature you will need an individual account. If you have one already please sign in.

Sign in.

Alternatively you can request an individual account here.