The STEM-designated MSBA program culminates with a spring analytics practicum building on students’ foundational training in the areas of data science, managing big data, machine learning, and data visualization. Student teams apply these skills to solve their sponsor firm's business problem using the firm’s proprietary data, possibly integrating with public datasets. Deliverables include a robust technical handover package with empirical models, code and cleansed data, a business deck explaining the benefits and methodology, and an executive dashboard developed on a visualization platform such as Tableau.
At the Nexus of Business, Data, and Technology
Data Savvy and Business Smart
Emory University Goizueta Business School delivers a STEM Master of Science in Business Analytics (MSBA) program. The MSBA is a full-time program for young professionals with limited work experience. It imparts strong technical and quantitative training plus comprehensive business acumen, all within a top business school.
Hands-on Student Learning
Tackling Your Big, Data-Driven Business Analytics Challenges
Student/Client Engagement
MSBA Program Managing Director
Scott Radcliffe
"The interaction between students an the client is a critical success factor for both sides. We ensure the students have solid consulting and project management processes to follow, but having a passionate, engaged client is the most critical element in the equation. It provides the students with the right learning environment and also gives them the required domain knowledge to generate real outcomes from their code and recommendations."
Analytics Practicum Project Sponsorship
How Does the Sponsorship Work
The ideal practicum has the following characteristics: 1) A good business question; 2) Rich, multi-source, accessible data; 3) Complex analytical requirements; 4) Strong sponsor engagement; 5) Desire to operationalize
Select Past Analytics Practicum Projects
| Sponsor | Business Problem | Project Description |
|---|---|---|
| Realtor.com | Identify fraudulent listings | At Realtor.com, we take the obligation to validate listings for fraud seriously. Our Listings Fidelity program applies Machine Learning techniques to evaluate new listings for their potential to put our consumers’ safety and finances at risk due to fraud. |
| Focus Brands | Moe’s Rewards churn reduction | Keeping existing customers is generally better for the business than trying to replace them. Predict customer churn likelihood to improve Moe’s Rewards retention efforts. |
| Best Buy | Reduce in-home service visits by enhancing phone problem resolution | Building intelligent assistants that aim to guide call center agents and technicians through the process of diagnosing and repairing problems with major appliances in our customers’ homes. |
| CONA Services – Technology Services for Coca-Cola Bottlers | In-store product assortment recommendation | Utilize in-store shelf image data and combine it with internal data like invoices, customer segmentation, scan volume to optimize correct product assortment at the store. |
| Truist Bank | Customer complaint analysis using NLP | Leveraging call center notes and topic algorithms to classify complaints for both prioritization of service capabilities and real-time routing. |
| Intercontinental Hotels Group | IHG Rewards Club Member churn | Predict likelihood of IHG Rewards Club members future trip behavior and guest preferences for personalized targeted communications. |
| The Home Depot | Space planning and visual recommendation | Develop a data-driven and modeling approach to categorize products and systematically recommend guidelines for visual design and space arrangement. |
| Stanley Black & Decker | Using digital customer feedback for product insights and sales volume forecasting | Using product commentary from retailer and influencer online presence apply NLP to reviews/ratings data and use these to make investment decisions on new products or update an existing product. |
Practicum Clients
Case Study: Emory Healthcare
-
The Challenge
Hospitals could improve post-operative care for all patients if they had an efficient, accurate, cost-effective way to identify the likelihood of infection among post-operative patients. -
The Approach
MSBA Capstone students partnered with Emory Healthcare to analyze a database of 6,482 patient records in which 317 had a wound infection and 227 had a systemic infection. They then trained a classifier with doctors’ notes as independent variables and the outcome label (infected or not) from NSQIP database as dependent variables, so that the classifier could learn to detect infection based on the doctors’ notes. -
The Outcome
The classifier developed by the Capstone students had a specificity of 0.92 and a sensitivity of 0.89 to classify whether each patient was infected after surgery. The implication is that post-operative patients can be labeled for infection and more closely monitored. Even better, Emory Healthcare can help doctors identify crucial symptoms related to infection and take steps to potentially reduce the infection rate.