33. Linear Regression in Python with sklearn
33.1 Why Predict a Number (Regression)
Why. Many India business questions are numbers: rent in ₹, daily pedha sales, delivery minutes. Classification says yes/no. Regression predicts a number.
What. Linear regression draws a straight-line style relationship: “as flat size goes up, rent tends to go up.” sklearn’s LinearRegression finds the best-fitting line (or plane) from your training rows.
How.
Spot a regression question
- Ask: “Is the answer a number I care about in ₹ / kg / minutes?”
- If yes → regression (this chapter). If yes/no or category → Chapter 34.
- Write one sentence: “I will predict ___ from ___.” Example: predict monthly rent from area_sqft and city tier (fictional Pune flats).
Worked example — fictional rent rows.
import pandas as pd
rent = pd.DataFrame({
"area_sqft": [450, 600, 750, 900, 1100],
"rent_inr": [12000, 15500, 19000, 22500, 27000],
})
print(rent)
print("Target y = rent_inr (a number). Feature X = area_sqft.")
What you should see. A tiny table where larger flats tend to cost more. Linear regression will learn that slope from data — not from a hand-typed formula.
Fictional data
Rents and flat sizes are teaching samples only, not real listings.
Ravindra Bagale's Tip
Regression = number. Classification = class. Interview madhe he ek line clear asa.