🔍 02 — EDA and Cleaning

First Checks

df.head()
df.shape
df.info()
df.isna().sum()
df["churn"].value_counts(normalize=True)

Cleaning Tasks

• standardize column names
• fix data types
• handle missing values
• remove exact duplicates
• inspect invalid values
• document decisions

EDA Questions

• What is the churn rate?
• Which customer segments churn most?
• How does tenure relate to churn?
• How does monthly spend relate to churn?
• Are there missing values related to churn?

Charts

• churn count plot
• numeric distributions
• box plots by churn
• correlation heatmap

---

Next

➡️ 03-feature-engineering