White Paper



Real-Time Analytics in Banking & Finance: Use Cases

By Seshika Fernando
Senior Technical Lead, WSO2

1. Introduction

Today, enterprises are looking for innovative ways to digitally transform their businesses - a crucial step forward to remain competitive and enhance profitability. There are key technology enablers that support an enterprise’s digital transformation efforts, including analytics. Real-time insights and data in motion via analytics helps organizations to gain the business intelligence they need for digital transformation. From a business perspective, the potential benefits it can offer an organization are many - you can use location and contextual data to create better customer experiences; create radically new data-based products for your business; make more informed decisions in complex scenarios; carry out effective monitoring and analysis; detect even the smallest change and trigger immediate action; and extend your solutions to analyze the past, present, and the future.

While these benefits are applicable to most organizations across diverse industries, a key advantage of analytics is that it can be customized to create solutions to meet the specific requirements of a particular industry.

This white paper will focus on the business benefits extended to the banking & finance industry and discuss some common use cases within this domain.

2. Analytics for Banking & Finance - An Overview

The world of banking & finance is a rich playground for real-time analytics. It has all the necessary ingredients; exploding data volumes, millisecond latencies, extreme volatilities and the need to detect complex patterns in real-time and act on them immediately. The ability to correlate, analyze and act on data, such as trading data, market prices, company updates, and other information coming through multiple sources at lightning speed is imperative to organizations within this industry.

2.1 Sample Use Cases

2.1.1 Money laundering/payment fraud detection

Money laundering detection and payment fraud detection are two important use cases in the financial industry. While the existence of both can not only inflict great financial loss, it could also cause significant damage to the respective bank’s corporate image. Unlike other industries, the corporate identity of a bank is critical to its existence and is a reflection on its credibility. Any sort of damage to its image could result in serious repercussions, even pushing the organization towards bankruptcy.

Streaming analytics offers comprehensive, real-time anomaly detection mechanisms to help banks and financial institutions to safeguard themselves from fraudulent activities. With streaming analytics, banks can easily convert their domain knowledge regarding fraudulent behavior to real time rules, use Markov modelling and Machine Learning to detect unknown abnormal behavior, and use scoring functions to reduce the number of false alarms being raised. Markov models are generally used to model randomly changing systems, and in the case of fraud detection, it helps to identify rare transaction sequences. This is especially useful in identifying complex fraudulent activity carried out not as one transaction but broken down into a series of smaller transactions by experienced crime rings. Machine learning enables computers to learn behavioural patterns on their own by referring to large amounts of past data without being explicitly programmed. Algorithms such as Clustering help a computer program to model ‘normal’ behavior by looking at past transaction trends. Therefore, this helps banks to identify new types of fraud by looking for transactions that differ from the normal behaviour that the machine learning algorithm has modelled.

Figure 1

Figure 1

For a more detailed account of these techniques, refer to Fraud Detection and Prevention: A Data Analytics Approach.

2.1.2 Risk management

In rapidly changing capital markets, it is no longer adequate to measure risk as an end of day process. Trading decisions can significantly alter exposures in a millisecond as traders with exposures to Bear Stearns found out the hard way in March 2008. In order to assess risks to market portfolios and take corrective measures in real-time, capital markets are now moving towards intra-day value at risk computations.

Streaming analytics can be leveraged to support these risk computations and aide banks to minimize and manage risk. With streaming analytics, banks can obtain a low latency, high-performance solution that listens to market prices as well as real-time changes to portfolios and compute value at risk on the fly. By employing risk calculations in a streaming fashion, financial institutions can stay several steps ahead of its competition by ensuring that portfolios are safe from intraday market fluctuations.

2.1.3 Stock market surveillance

Unethical profit gain via artificially inflating or deflating stock prices, exploiting prior knowledge of company proceedings, advance knowledge of impending orders, and insider trading are common forms of stock market manipulation. And to prevent these, a stock exchange can incorporate streaming analytics into their overall surveillance efforts.

Streaming analytics is a great stock market surveillance tool that can spot even the mildest form of market manipulation, ranging from insider trading to price manipulations for profit gain in real time. Even though stock market manipulation is considered illegal worldwide, identifying suspicious behavior is often rendered cumbersome or impractical due to the volume and velocity at which trading is executed. Thus, a majority of illegal trading activities are not captured as and when they occur. By joining market data feeds with external data streams, such as company announcements, news feeds, Twitter streams, etc., streaming analytics can instantly identify activities that are possible attempts of market manipulation. By doing so, regulators can be alerted in real time so they can take early action, even before the manipulation takes place.

Figure 2

Figure 2

Refer to our latest case study where WSO2 built a real-time stock market surveillance tool for the Colombo Stock Exchange.

3. Conclusion

Analytics used to be a term reserved for data scientists - a word heard by many, but understood by a few. This is no longer the case. Enterprises that do not reap the benefits of analytics will soon be edged out by their competitors. With data being a key component in any business today, enterprises are forced to look for new ways to analyze this data and gain insights into their business. Thus, in today’s business world, analytics has become vital to improve customer experience, increase market reach, optimize budget spend, enhance business processes, and find and eliminate anomalies. All of these eventually translate to improved revenue for any business.

It’s clear that streaming analytics is widely applicable within the banking & finance industry, helping organizations to get a better grasp of current trends, secure portfolios from adverse market effects, and safeguard investors from unscrupulous behaviour of fraudsters.

It can also be used for specific solutions and use cases in other industries as well. Refer to our white papers that cover other industry solutions for more details:

  • Retail
  • Fleet Management
  • Smart Energy Analytics
  • Social Media Analytics
  • QoS Enablement
  • System and Network Monitoring

With the pace at which the world is transacting, analytics that are computed as batches will no longer be relevant. The current need is to perform complex analytics in real-time so enterprises can act on them before the opportunity goes by. Streaming analytics is a perfect fit for this role as it can receive multiple types of data from multiple sources, correlate them, process them, and provide meaningful insights all in a matter of milliseconds. Irrespective of the industry, streaming analytics can create a winning strategy for your business.

If you want to try out these ideas, please checkout WSO2 Stream Processor. WSO2 Stream Processor (WSO2 SP) is an open source stream processing platform. It can ingest data from Kafka, HTTP requests, message brokers and you can query data stream using a “Streaming SQL” language. With just two commodity servers it can provide high availability and can handle 100K+ TPS throughput. It can scale up to millions of TPS on top of Kafka.

If you found these use cases helpful and/or applicable to your organization or have similar use cases, we’re happy to further discuss your requirements and take you through a demo. Please contact us and we’ll get in touch.

For more details about our solutions or to discuss a specific requirement


Interested in similar content?