Your Resource for All Things Apps, Ops, and Infrastructure

Applying Data Science for Financial Services

The financial services industry has been experiencing a technology-driven shake-up over the last several years. Customers’ increasing reliance on mobile apps for banking, investment, and payment needs has been a major force in the digital transformation of an industry that has traditionally relied on brick-and-mortar service delivery. However, the industry seems to be embracing this disruption.

Due to the growing digital trends, there has been an influx of data available to financial organizations. We are finding the industry’s focus on cloud-based data analytics is maturing with the demand for our Financial Data Hub increasing. Further, published reports, like this whitepaper from Dataiku, have found that many financial service organizations are beginning to look ahead at emerging technologies, like data science, blockchain, and the adoption of artificial intelligence (AI).

There’s no question that financial service organizations have access to copious amounts of data. These emerging technologies allow organizations to leverage more of their data than ever before to drive better insights into their biggest business challenges, including fraud detection, risk management, churn reduction, regulatory reporting, and new customer segment identification.

Data Science Critical for Financial Services

Particularly, data science has become vital for organizations in the financial industry, including banks, credit unions, stock brokerages, accountancy, and insurance companies. Data science becomes imperative when a company has a large quantity of data – often of many different data types. The rapid growth in data sources, quantity, and types – including transactional, IoT, and customer data – has required financial organizations to leverage data science to manage the vast amounts of diverse data in a way that allows them to gain better customer insights at scale.

Data science isn’t totally new to financial services though. Predating data science, quantitative analysts developed complex statistical models that financial organizations would rely on for decision-making on topics like pricing, investments, and risk. The application of data science emerged from the industry’s use of quantitative analysts, which is part of the reason why the financial industry is generally further along in the adoption of data science and AI.

A modern data science practice gives financial organizations the ability to leverage larger and more complex datasets for deeper decision-making across the entire enterprise.

Many banks and other financial companies have found themselves struggling to mature into a modern data science practice. We found the key challenges our financial service clients face are the management of the influx of data, an overwhelming morass of analytic models, model versions and interconnections, and the operationalization of data science in a way that allows the insights to be translated into action throughout the entire organization.

To overcome these challenges, we recommend using Dataiku who provides a leading collaborative data science platform called their Data Science Studio (DSS), which enables self-service analytics that operationalizes machine learning. Vertical Trail utilizes Dataiku DSS as the central tooling to support the core requirements of our clients’ data science practice. We find their platform to have many benefits for our clients, particularly those in financial services.

1. Data Management

Dataiku DSS facilitates automated and secure data preparation that includes gathering, cleansing, formatting, and merging in a cost-effective manner.

For data gathering, the tool has the ability to gather data through secure connections to multiple sources: Hadoop data sources, enterprise SQL databases, and Excel spreadsheets. More so, you can reuse preexisting data types in the platform. This is particularly important for financial services organizations that have many different data types, including data for regulatory standards – including customer transactions and conversation transcripts as well as unregulated data – including data from marketing campaigns and web.

For data cleansing, DSS utilizes a drag-and-drop, point-and-click interface allowing visualization of workflows, which can be useful when presenting to the team. Ultimately, the highly intuitive environment creates a uniformed dataset that can be acted on for advanced analytics and AI for better decision-making.

2. Advanced Analytics

The Data Science Studio’s data management functionality enables organizations to properly utilize the tool’s robust analytics functions. Dataiku DSS gives users complete control and flexibility of the algorithms used as part of your organization’s analytics solutions.

Dataiku leverages many ML libraries, including Spark Mlib and the SciKit machine learning package in Python. This provides guided analyses with instant feedback on model performance. However, you are not just limited to their available libraries – the platform also allows you to use custom algorithms that can be reused or merged together. This allows financial organizations to build diverse predictive analytics solutions that take their uniformed raw dataset and transform them into business-impacting services that solve a variety of financial use cases including fraud detection, churn reduction, and new customer identification.

3. Organization-Wide Use

Dataiku DSS provides organization-wide benefits. The data management and advanced analytics functionalities outlined above enable business analysts, data scientists, and engineers to collaborate on projects and operate more efficiently. The platform is particularly critical for larger organizations that consist of both large technical and business teams, as many financial institutions do. We find it to be key for business users to interact with their technical teams throughout the development cycle.

Additionally, the tool’s visualization capabilities are useful in presenting all scopes of a project, including workflows, graphs, and statistics to business stakeholders. This helps stakeholders get a better sense of AI projects from start to finish and often improves buy-in to increase the organization’s AI spend. For financial organizations, another organizational use of Dataiku is the ability to present all the details of projects, including data use and results, to regulators. By having organizational involvement in data science projects, financial institutions will find they will more effectively apply data science and AI at scale to solve some of their organizations’ biggest challenges.

Optimizing Data Science

Vertical Trail accelerates the implementation and speed to scale of Dataiku DSS as part of our clients’ data science practice. We specialize in the underlying technologies necessary to optimize DSS deployments. Most importantly, we implement our financial clients’ entire environments with integrated data security that meet industry regulatory standards. We also provide ongoing support for Dataiku with Big Data and Analytics Managed Services.

Subscribe to the AHEAD i/o Newsletter