Course contentsShow
← Data Science
Lesson 480 of 2,14513. Correlation AnalysisFree lesson

Scatterplots and Visual Assessment

Using scatterplots to visually inspect correlation before calculating r and identifying patterns.

Scatterplots and Visual Assessment

What You'll Learn

You'll learn how to use scatterplots to visually inspect the relationship between two variables before calculating Pearson's r, helping you spot patterns and potential issues.

Why Look Before You Calculate?

Before computing any correlation coefficient, always plot your data first. Numbers like Pearson's r can be misleading if you don't understand the underlying pattern. A scatterplot shows you the raw relationship between two variables, revealing whether correlation analysis even makes sense.

Creating a Mental Picture

Think of a scatterplot like looking at footprints in sand before measuring the distance between them. You want to see:

  • Direction: Do points trend upward (positive) or downward (negative)?
  • Strength: Are points tightly clustered along a line, or scattered widely?
  • Form: Is the relationship linear (straight-line pattern) or curved?
  • Outliers: Are there unusual points far from the main pattern?

What to Look For

When you examine a scatterplot, you're checking if Pearson correlation is appropriate:

  1. Linear pattern: Points should roughly follow a straight line, not a curve
  2. No extreme outliers: A single outlier can dramatically distort r
  3. Consistent spread: The scatter shouldn't fan out or compress at one end
  4. Adequate range: Data should span a reasonable range of values

If you see a clear curve, clusters, or other non-linear patterns, Pearson's r might give you a misleading impression of "no correlation" even when a strong relationship exists—just not a linear one.

Key Takeaway: Always create and examine a scatterplot before calculating Pearson's r to verify the relationship is linear, identify outliers, and ensure correlation analysis is appropriate for your data.