Schedule a Demo

Business Analytics

Guest Post – Knowledgent’s TeKathon Using Trifacta for Financial Services Analytics

< Back to Blog
 
August 26, 2015

Marjon Rahimian, Informationist at Knowledgent, blogs on the Financial Services team of this year’s TeKathon II. Knowledgent analytics firm had two teams – one Health and Life Sciences and another Financial – learn Trifacta software in order to challenge them to a data wrangling contest. With the added challenges of real-world constraints of time and resources, each team’s data sets were then assessed based on innovation, execution, business model and presentation. Read about how their opponents on the Health team faired on our previous blog here

As a wrap on TeKathon II comes to a close, we delve into the minds of the Financial Services and Insurance team, who took home the prestigious TeK-Cup, on the particulars of their winning use case…

In today’s fast-paced, technology-driven landscape, consumers have come to expect far more for less in all aspects of their lives, including personal finances.

The ascendance of robo-advisors – automated, model-based online services that provide investment recommendations – poses an increasing threat to our traditional Financial Service clients by claiming to provide faster, cheaper and more transparent investment solutions powered by sophisticated technology and minimal to no human intervention.

In order to effectively compete, our clients need to find new ways of differentiating themselves or else they run the risk of losing significant market share to these emerging automated services.

Since automation can never truly replace certain aspects of human expertise and interaction – particularly when it comes to dealing with personal finances – Financial Service firms must capitalize by offering a solution that combines the best of both worlds.  Knowledgent helps its clients build these cutting-edge solutions by augmenting the expertise of financial advisors with the advanced technological capabilities of a robo-advisor platform to address the demand for smarter, more personalized investment advice driven by sophisticated data analytics. This hybrid approach has come to be known by some as a “robo-assisted advisor.”

A lot goes on behind the scenes to ensure that a robo-assisted advisor can provide his or her customers with high quality custom recommendations. One of the major challenges they face is making sure that the data acquired from a broad range of diverse sources is clean, complete, and organized prior to delivering analytics and visualization to their customers. For TeKathon II, we used Trifacta’s data-wrangling capabilities to solve for this challenge and to demonstrate its value contribution to our robo-advisor business case. Trifacta’s ability to discover, cleanse, and link disparate datasets helped ensure that the data being used to generate customer recommendations was both comprehensive and accurate – before being analyzed and visualized.

For TeKathon II, we went through the series of steps outlined below to delve into Trifacta’s capabilities and showcase its value contribution to our business case.

  • 1.) Identify and Create Robo Advice Datasets: It was difficult to obtain pre-existing data that was relevant to our business case, so we created datasets from scratch that simulated real-life robo-advice data. Two of the datasets housed customer demographic and finanical information to paint a full 360° view of each customer, while the rest comprised of disparate data around the different asset class types to help fuel the investment recommendations.
  • 2.) Transform Datasets from Clean to Messy: To simulate the messiness of real-world data, we populated the datasets with various text, numerical, and formatting errors, then documented all errors to evaluate Trifacta’s ability to identify and rectify each error.
  • 3.) Delve into Trifacta’s Data-Wrangling Capabilities: We loaded datasets into Trifacta to discover, cleanse, and link all the datasets using common identifiers. The end result was a single, merged dataset that housed all demographic, financial, current and recommended investment information for each customer.
  • 4.) Add Rules to Customize Investment Recommendations: We applied comprehensive business rules (simulating sophisticated investment algorithms) to the data based on each customers demographic and financial inputs as a means of customizing their recommended investments.
  • 5.) Showcase Investment Recommendations via Dashboard Visuals: We selected three customer examples and built custom dashboards in Tableau for each one to showcase how investment recommendations would typically be displayed and explored in a real-life scenario.

Using a hands-on approach to delve into a tool like Trifacta made it all the more evident that data-wrangling is an essential complement to any form of analytics and visualization. We believe this to be specifically true for robo-advice, but also applies more generally to the type of data exploration and transformation that is increasingly required with the proliferation of big data and analytics.

As the volume, variety, and velocity of data continues to grow, data wrangling will become increasingly important to ensure veracity.  Advanced tools like Trifacta can help automate or accelerate much of the laborious work associated with data preparation by providing organizations with a highly intuitive and easy-to-use visual interface, which when used appropriately can be a key differentiator for data scientists and business analysts to extract value from data assets.

To see more photos from TeKathon II, check them out on our Flickr page!

If you’d like to see how you can wrangle your own data, check out our resources section to learn more about the Trifacta solution and sign up for Trifacta Wrangler