Aus Daten werden Entscheidungen.

# Umsatzprognose je Region — DATALAB
import pandas as pd
from sklearn.ensemble import GradientBoostingRegressor

def train_forecast(df: pd.DataFrame) -> GradientBoostingRegressor:
    features = ["region", "quarter", "marketing_spend"]
    X, y = df[features], df["revenue"]
    model = GradientBoostingRegressor(
        n_estimators=400, learning_rate=0.05, max_depth=3,
    )
    model.fit(X, y)
    return model  # R² 0.986 auf Holdout

-- Marge je Kunde, rollierend 12 Monate
SELECT
    c.customer_id,
    c.name,
    SUM(o.revenue)            AS revenue,
    SUM(o.revenue - o.cost)   AS margin,
    ROUND(AVG(o.margin_pct), 1) AS avg_margin_pct
FROM analytics.orders   o
JOIN analytics.customers c ON c.customer_id = o.customer_id
WHERE o.order_date >= DATEADD('month', -12, CURRENT_DATE)
GROUP BY 1, 2
ORDER BY margin DESC;