32. Features and Labels in Machine Learning

32.4 Building a Tiny Feature Table in pandas

By Ravindra Bagale · Last updated: 24 October 2026

Why. Your hands need one tiny pandas build before sklearn models arrive.

What. Build X and y from a clean fictional table.

How.

import pandas as pd

raw = pd.DataFrame({
 "order_id": ["BLK-1", "BLK-2", "BLK-3", "BLK-4", "BLK-5", "BLK-6"],
 "city": ["Pune", "Solapur", "Pune", "Nashik", "Kolhapur", "Nagpur"],
 "items": [2, 6, 3, 5, 4, 8],
 "distance_km": [1.1, 7.5, 2.0, 3.4, 4.2, 5.1],
 "rain_mm": [0, 10, 0, 0, 2, 0],
 "was_late": [0, 1, 0, 0, 1, 1],
})

feature_cols = ["items", "distance_km", "rain_mm"]
X = raw[feature_cols].copy()
y = raw["was_late"].copy()
print(X)
print(y.value_counts())

What you should see. X has only allowed inputs. y is the 0/1 label. order_id and city stayed out for this tiny numeric-only demo (city can return later as a category feature).