pipe operator in dataflows

Pipes (df.pipe) turns a DataFrame into the “noun” that flows through small, focused “verb” functions. Instead of mutating globals across cells, you build a linear, readable pipeline.

# helper processing functions
def filter_active_users(df):
    return df[df["events"] > 3]
 
def drop_bots(df):
    return df[~df["user_agent"].str.contains("bot", case=False)]
 
def sessionize(df, timeout_minutes: int):
    ...
    return df_with_sessions
 
# pipeline
def clean_events(df):
    return (
        df
        .pipe(filter_active_users)
        .pipe(drop_bots)
        .pipe(sessionize, timeout_minutes=30)
    )
 

Combining with df.query one can create super clean pipelines, while allowing for testing and reusability.