Lesson ProgressPhase 6 of 6
Phase 6Closing
Closing: Outliers and Data Quality

Consolidate learning and connect to broader unit goals and real-world applications

Your Data Detective Toolkit

You've completed your first data quality investigation. Sarah is confident you can now handle messy real-world data with the same rigor professional analysts use.

What You Can Now Do

Statistical Tools

  • • Calculate z-scores: z = (x - μ) / σ
  • • Apply the ±2 rule for outlier detection
  • • Compare statistics with and without outliers
  • • Measure impact on planning calculations

Business Judgment

  • • Distinguish errors from real events
  • • Decide: keep, flag, or correct outliers
  • • Document decisions with reasoning
  • • Explain choices to stakeholders
How This Connects to the Café's Goal

Sarah can now build her forecasting models on clean, reliable data. The café's waste reduction plan depends on accurate predictions - and accurate predictions depend on clean data.

Today

Clean data + identified outliers

Next Lesson

Use clean data to predict future sales

Unit Goal

Reduce waste from 8-12% to 3%

When Will You Use This?

Any Data Analysis Job

Every dataset has outliers. Banks flag suspicious transactions, healthcare identifies unusual test results, retailers spot异常的 purchase patterns.

Business Consulting

When you present numbers to clients, they will ask "why is this different from the raw data?" You need documentation and reasoning for every decision.

Research & Science

Every scientific study begins with data cleaning. Your outlier detection skills are foundational to any research career.

Quality Control

Manufacturing uses the same ±2 rule to identify defective products. The same statistical thinking applies everywhere.

Outlier Detection & Data Quality - Lesson 3 Reflection
Reflect on your learning journey and growth in the CAP framework
0/3 Complete
CONFIDENCE
Outlier detection requires making judgment calls that aren't always clear-cut. Describe a time in this lesson when you felt unsure whether a value was an error or legitimate business. How did you build confidence in your decision?
0 characters
🌊ADAPTABILITY
The z-score told us the $0.05 was "normal" mathematically, but you still identified it as an error. How did you adapt your approach to combine statistical rules with practical business judgment?
0 characters
PERSISTENCE
Cleaning data takes patience - you can't just delete everything that looks unusual. Describe a moment when you had to persist through the careful thinking required to make good data quality decisions.
0 characters
Progress: 0/3 reflections completed
Coming Up Next

With clean data in hand, Sarah can now build forecasting models to predict future sales. In the next lesson, you'll learn how to use the patterns in past data to predict what might happen next - the foundation of every business plan.