GroupBy & Aggregation

GroupBy & Aggregation

The split-apply-combine pattern is the workhorse of data analysis.

# Split → Apply → Combine
df.groupby("major")["gpa"].mean()

Multiple Aggregations

df.groupby("major").agg({
    "gpa":      ["mean", "max", "min"],
    "projects": "sum"
})

Sorting

df.sort_values("gpa", ascending=False)
df.sort_values(["major", "gpa"], ascending=[True, False])

Value Counts

df["major"].value_counts()  # frequency of each major

Real-world use: This is equivalent to SQL's GROUP BY ... HAVING — you'll use it in every data analysis project.