df["unique_products"] = df.groupby("customer_id").agg({"products": "unique"})

# Transform each element -> row - Pandas >= 0.25
df["unique_products"].explode()