Read Patterns In Data

What you will learn

• Understand that almost every useful insight from data fits into one of three basic pattern shapes (trend, comparison, anomaly), and that spotting which kind of pattern you are looking at is the first skill of reading data.
• Understand the difference between signal (a real pattern that means something) and noise (random variation that means nothing), and learn to ask whether a pattern is bigger than the noise before trusting it.
• Learn to read a simple chart honestly by asking two specific questions (compared to what, and is this bigger than the noise), and by describing what the chart shows rather than what you hoped it would show.
• Combine the three pattern shapes, the signal-noise discipline, and the honest chart reading into a one page Pattern Report on a real small dataset the student has access to.

What happens in this lesson

1. Chapter 1: Three Kinds Of Pattern

   Understand that almost every useful insight from data fits into one of three basic pattern shapes (trend, comparison, anomaly), and that spotting which kind of pattern you are looking at is the first skill of reading data.

   What you make: A saved pattern spotting note identifying one trend, one comparison, and one anomaly in real or imagined data.

2. Chapter 2: Signal And Noise

   Understand the difference between signal (a real pattern that means something) and noise (random variation that means nothing), and learn to ask whether a pattern is bigger than the noise before trusting it.

   What you make: A saved signal check note analysing one real pattern to decide whether it is signal or just noise, with specific reasoning.

3. Chapter 3: Read A Chart Honestly

   Learn to read a simple chart honestly by asking two specific questions (compared to what, and is this bigger than the noise), and by describing what the chart shows rather than what you hoped it would show.

   What you make: A saved honest reading note analysing one specific chart in plain language, naming what it actually shows and what it does not show.

4. Chapter 4: Your Pattern Report

   Combine the three pattern shapes, the signal-noise discipline, and the honest chart reading into a one page Pattern Report on a real small dataset the student has access to.

   What you make: A saved Pattern Report on a real small dataset showing at least one trend, one comparison, and one anomaly, with a short honest interpretation of what the data actually means.

5. Chapter 5: Read A Messy Dataset

   Apply pattern reading to a real dataset with more noise, missing context, and competing explanations, and decide which patterns are strong enough to trust.

   What you make: A saved evidence brief that identifies the strongest pattern in a messy dataset, explains why it is trustworthy, and names at least one alternative explanation.