Overview of Student Spotlight: Aaron Payne, Data Analyst
This episode of Data Skeptic features Kyle Polich’s conversation with Aaron Payne, an MBA student at Georgia Tech and newly appointed Senior Insights Analyst at Chick-fil-A. The discussion centers on Aaron’s career path from technical tax transformation into analytics, the value of business analytics across industries, and a real-world forecasting project he completed for a Colombian social services company. The episode highlights how analytics can drive better operational decisions and ultimately improve services for real people.
Aaron Payne’s Background and Career Path
Aaron shares a progression that reflects a blend of technical and business-oriented work:
- Started at Ernst & Young in technical tax transformation
- Learned Python scripting and data visualization
- Moved to Atrium Hospitality
- Worked across operations, commercial, HR, and finance
- Did ad hoc analytics, visualization, and some machine learning
- Began an MBA at Georgia Tech’s Scheller College of Business
- Studying business analytics
- Recently joined Chick-fil-A as a Senior Insights Analyst
- Will focus on internal consulting for analytics and supply chain
Aaron emphasizes that his goal has been to keep growing in analytics while staying connected to Atlanta.
Key Themes: Turning Data Into Insights
A major theme of the conversation is the difference between simply producing analytics and turning them into action:
- Aaron notes that Chick-fil-A intentionally uses terms like:
- Insights analyst instead of data analyst
- Decision science instead of data science
- The point is not just to build models, but to:
- Translate data into actionable decisions
- Support operational excellence
- Improve outcomes for customers and communities
This framing reflects his broader belief that analytics should have a direct business and human impact.
Georgia Tech Practicum Project: Forecasting for Confama
One of the most substantial parts of the interview covers Aaron’s practicum project at Georgia Tech.
Project Context
- Worked with Confama, a social services organization in Colombia
- Confama serves an affiliated population that receives benefits such as:
- Education
- Afterschool care
- Other social services
- They needed help forecasting future demand/population levels amid:
- Pandemic-related disruption
- Economic instability
- Urbanization and demographic shifts
Why It Mattered
The forecasting work had real operational consequences:
- Better forecasts help Confama plan resources more accurately
- Inaccurate forecasts could mean people don’t receive essential services
- The project highlighted the real-world stakes of analytics in a social-services setting
Data Challenges and Cleaning Work
Aaron describes the project as a realistic example of how messy applied data science can be:
- The data was not in English, which added translation complexity
- Some values were manually entered and contained errors
- There were data gaps and potential outliers
- The team had to consider:
- Seasonality
- Trend effects
- COVID-related disruption
- Other economic variables relevant to Colombia
He reinforces the classic principle: “garbage in, garbage out.” Data quality was essential before any modeling could be trustworthy.
Modeling Approach and Methodology
The team chose a modeling strategy based on both predictive performance and interpretability.
Models Considered
- Started with ARIMA, a standard forecasting method
- Expanded to SARIMAX:
- Seasonal ARIMA with exogenous variables
- Added a machine learning component using XGBoost
- Also explored other methods, including:
- Exponential smoothing
- Prophet
Why SARIMAX Worked Well
SARIMAX allowed the team to incorporate external economic indicators from Colombia’s national statistics bureau (DANE) such as:
- Employment rate
- GDP-related indicators
- Other sector-level economic variables
This helped account for factors outside Confama’s own historical data.
Why an Ensemble Model
Aaron explains that the final solution was an ensemble:
- Combined the strengths of statistical forecasting and machine learning
- Used RMSE weighting to combine predictions
- Reduced residual error and improved forecasting performance
The team also checked for:
- Multicollinearity
- Model interpretability
- Prediction intervals acceptable to the client
Lessons About Forecasting in the Real World
Aaron stresses that model selection is not purely technical — it depends on the stakeholder and the business setting:
- Confama needed a model they could understand
- The forecast had to fit within a one-semester academic timeline
- The model needed to reflect the organization’s existing expertise and business context
He frames this as a mix of art and science:
- Test multiple approaches
- Use expert guidance
- Balance accuracy, interpretability, and practicality
Handling COVID and Other Disruptions
The conversation briefly touches on whether COVID-era data should be discarded as an outlier.
Aaron’s response is nuanced:
- In some contexts, you might exclude COVID data
- But for Confama, disruption is part of the normal operating environment
- Colombia has experienced other major shocks, including government changes and economic volatility
Instead of removing the period, the team created a COVID indicator variable so the model could account for economic disruption in a structured way. This made the forecast more relevant to Confama’s reality.
Chick-fil-A’s Analytics Culture and Values
Aaron also reflects on his new role at Chick-fil-A and the company’s broader aspirations.
Notable Takeaways
- Chick-fil-A aims to be “one of the most caring companies in the world” by 2030
- Analytics is used to support:
- Supply chain decisions
- Environmental responsibility
- Operational efficiency
- Customer delight
- Aaron is impressed by how the company ties analytics to stewardship and care, not just cost reduction
He sees this as a strong example of values-driven data work.
Advice for Working While Studying
Aaron offers practical advice for anyone balancing work and an advanced degree:
- Plan ahead
- Don’t go into it blindly
- Build a network
- Learn from people who have already done it
- Recognize that balancing school, work, and life is hard
His personal advice:
- If possible, do it before major life commitments
- Expect sacrifice, late nights, and fatigue
- A well-thought-out plan makes the process much more manageable
What’s Next for Aaron
Aaron closes by describing his future goals:
- Continue at Chick-fil-A
- Move deeper into supply chain analytics
- Finish his MBA at Georgia Tech
- Grow into a role that bridges:
- Business decisions
- Data science
- Machine learning / AI
- He is especially interested in:
- Agentic AI
- More advanced data science workflows
- Long-term growth in the analytics space
His end goal is a role that combines business understanding with rigorous technical modeling.
Final Thoughts
This episode is a strong example of:
- How analytics education translates into real-world business impact
- Why forecasting is as much about context and communication as it is about models
- How data science can support social good, operational excellence, and better decision-making
Aaron’s story shows a practical path from student to industry analyst, with a clear emphasis on interpretability, responsibility, and using data to help people.