How Data Training Brings Employee Impact in Step with Enterprise Mission and Objectives

Upskilling the Data Analytics Workforce

machine learning
Discover Financial Services
machine learning
2020-Ongoing
machine learning
Financial Services
Discover Financial Services logo over New York skyscrapers

Synopsis

KPI:
Successful completion of capstone projects that produced tangible, data-driven results for the company.
Project length:
2020-Ongoing
Employees trained:
TBD

At a Glance

Sustaining robust data governance guidelines and ensuring that your staff feels comfortable leveraging data within your systems is a challenge for even the most sophisticated of organizations. A Fortune 500 company with an expansive asset portfolio, Discover Financial Services (DFS) sought expert guidance to help onboard new employees so they could effectively navigate the company’s systems and data tools.

Discover engaged Data Society to design a customized two-week onboarding data boot camp that was in line with the mission and objectives of the organization. The result was a well-received and relevant program that Data Society delivered in a virtual classroom during the pandemic lockdown and a broad data science training strategy that met the varied needs of the new DFS workforce.

Client Profile

Bio:
Discover Financial Services is an American financial services company that owns and operates Discover Bank, provider of checking, savings, and credit card accounts, as well as personal, home equity, and student loans.
Industry:
Headquarters:
Financial services
Riverwoods, Il
Revenue:
$10.5B
Founded:
1985
Size:
15,549 Employees

The Challenge

Discover Financial Services (DFS) has developed a robust data ecosystem unique to their needs and objectives which has helped them maintain their top standing in the financial industry. However, as the organization has continued to grow, leaders at DFS identified a knowledge gap between new employees and existing staff that made it difficult for the new hires to fulfill their tasks effectively. Discover selected Data Society to design a training roadmap for incoming analysts to understand the organization's data environment and have a command of R programming fundamentals and its applications to financial services. Data Society developed a custom technical onboarding training for all new employees to equip them for success in their many functions across the organization. 

Current State

DFS onboards analysts with a range of backgrounds: recent graduates to seasoned financial data scientists. With a variety of skills to shepherd into the organization, the L&D department has the unique challenge of ensuring that new hires, regardless of background, can operate and succeed within their business functions. 

DFS, like any other large organization, operates within a nuanced set of tools and procedures: data catalogs, governance, processes, and methods to analyze portfolios, risks, credit, and loans. New employees need to understand all of these techniques. Even seasoned analysts require an understanding of the specific data landscape in which they will operate; this takes thoughtful onboarding, which Data Society was able to provide.

Training Objective

DFS's 2020 strategy includes a mission to become increasingly data-informed as a business. Data Society worked with DFS to create a tailored 40-hour data science training program for employees using R, SQL, and Snowflake, aimed at upskilling employees and working to fulfill the data-centered vision.

Overcoming a Habit

DFS has been working to modernize its methods. While it's tempting to stick to the tried and true legacy software, free open-source tools like R programming have been paving the way to more affordable and sustainable modeling environments that enhance analysts’ powerful capabilities and give company leaders more impactful insights on how to drive the organization forward. 

The Solution

Data Society delivered the stepping-stone course in DFS’ "Discovering Analytics” pathway. The program was designed collaboratively by data scientists, instructional designers, project managers, and key stakeholders on both DFS and Data Society teams. Roundtable sessions were the catalyst for understanding how the organization is currently using data to measure default risk, forecast portfolio growth, evaluate credit models, and more. The shift in the vision for data science at DFS was a centerpiece of the discussion, along with updating legacy practices to incorporate the more advanced techniques and tools.
Overcoming COVID-19

COVID-19 complicated the initial face-to-face delivery plan, and Data Society quickly adapted to a virtual instructor-led schedule. Data Society worked with DFS L&D managers and IT staff to create a virtual learning environment that would offer students remote access to all training materials and tools.

The platform we developed required no downloads onto work computers, allowing us to proceed with the course despite the students' loss of access to the learning lab. Project managers and instructional designers worked together across the two teams to create a DFS specific training environment, powered by AWS, RStudio Cloud, and Zoom to create online synchronous training that provided an engaging learning environment.

Learning Objectives
  • Manage relational databases and tools
  • Learn to access Discover’s data with SQL
  • Manipulate, aggregate, and reorder data with SQL 
  • Explain what Data Science is, how it is used at Discover, and which skills a data scientist needs
  • Import, export, clean and manipulate data in R
  • Set up a workflow pipeline using R 
  • Mine data to find latent patterns and
  • Evaluate the effectiveness of clustering analyses
  • Understand the purpose and implications of what clustering methods can and cannot achieve 
  • Identify use cases where clustering analyses are relevant, and where they are not applicable
  • Identify use cases for predictive analytics
  • Build classification models to anticipate events
  • Evaluate the accuracy of predictive algorithms 
  • Tune algorithms to optimize performance 

The Results

67

%

Pre-course Average Baseline Score

86

%

Post-course Average Testing Score

28

%

Improvement from Training

Pre-assessments allowed students with specific needs to enroll as part of the course; Data Society selected learning objectives relevant to student functions. That same survey established each student's current skill level and set a benchmark to measure against in a post-course skills evaluation. The result of the first cohort was a statistically significant 28% improvement in technical knowledge and demonstrable progress towards the pre-set learning objectives defined in the early engagement. 

Hard Skills Acquired 
  • Work in Snowflake and SQL
  • Use foundational techniques R
  • Work with data in R 
  • Manipulate and aggregate data
  • Use clustering techniques
  • Cluster categorical data
  • Deploy hierarchical clustering
  • Create supervised ML models
  • Write classification algorithms
  • Create advanced models 
  • Optimize models 
Building a Lasting Relationship

Data Society and DFS will continue to deliver the onboarding program and explore more advanced pathways for learners to acquire abilities in advanced forecasting, machine learning, and deep learning. This partnership will also provide banking analysts with the latest financial data science resources to help DFS stay competitive and agile in a demanding financial industry. 

cross linkedin facebook pinterest youtube rss twitter instagram facebook-blank rss-blank linkedin-blank pinterest youtube twitter instagram