Review previous learning and connect to today's application activities
Sarah has analyzed the café's weekend data and calculated that the average (mean) transaction is $12.50. She's ready to use this number to help the café predict revenue and plan inventory.
But there's a problem. When she looks more closely at individual transactions, something doesn't add up.
Sarah calculates: Average = $12.50
But then she notices these transactions:
- Coffee: $4.25
- Muffin: $2.75
- Latte: $5.25
- Lunch combo: $12.95
- Catering Order: $127.50
- Data Entry: $0.05
The Friction Point
If most transactions are $3-15, how can the average be $12.50? And more importantly - what should Sarah do with those unusual values? Are they errors? Legitimate business events? Something in between?
This is where outlier detection becomes essential for every data analyst.
🔍 Detect Outliers
Use z-scores to objectively measure how unusual each data point is
⚖️ Make Quality Decisions
Decide whether to keep, flag, or remove outliers based on business context
💼 Explain Your Reasoning
Defend data cleaning decisions to café management with evidence
📊 Improve Analysis Accuracy
Ensure forecasts and recommendations are based on reliable data