Skip to content
Data Science & Analytics

Exploratory Data Analysis

Correlation and Its Traps

8 min+75 XPStep 1/10

Moving together ≠ causing

Correlation measures whether two variables move together, on a scale from -1 (perfectly opposite) through 0 (no linear relationship) to +1 (perfectly in sync). Ice cream sales and drowning deaths correlate. So do ice cream sales and sunscreen sales.

The trap everyone knows and still falls into: correlation is not causation. Two correlated variables might be:

  • A causing B — more support tickets really do drive more churn
  • B causing A — maybe failing accounts generate tickets instead
  • C causing both — summer heat drives both ice cream and swimming
  • Coincidence — with enough columns, some pair will correlate by pure luck

analogy

Correlation is two people dancing in sync. It doesn't tell you who's leading — or whether a DJ is telling both what to do.