Data Preprocessing — ML Technical Interview Questions & Answers

Cleaning, encoding and scaling data before training.

Data Preprocessing

  • Simple meaning: Raw sheets are messy. Preprocessing is cleaning before train: fill or drop missing marks, fix “Pass ” with a space, turn city names into numbers, scale hours and income so one column does not dominate.
  • Picture: Column hours is 1–10. Column fees is 20,000–80,000. KNN will think fees is “more important” only because the numbers are bigger. Scale both (StandardScaler / MinMax) on train, then apply the same scaler to test.
  • Typical steps: drop junk rows → handle NA → encode categories (one-hot or label) → scale numeric columns → then fit the model.
  • Trap: fitting the scaler on train+test together (leak). Trap: one-hot encoding after the split in a way that test gets new columns the model never saw — plan categories on train.

Technical Interview Tip: Clean → encode → scale on train. Transform test with the same fit.

Code Examples for Data Preprocessing

Data Preprocessing — working example

xs = [10, 20, 30, 40]
mean = sum(xs) / len(xs)
var = sum((x - mean) ** 2 for x in xs) / len(xs)
std = var ** 0.5
z = [(x - mean) / std for x in xs]
lo, hi = min(xs), max(xs)
minmax = [(x - lo) / (hi - lo) for x in xs]
print([round(v, 3) for v in z])
print(minmax)
# sklearn.preprocessing.StandardScaler fits mean/std on train only

Data Preprocessing — scaling keeps features comparable — critical for distance-based models and many optimizers.

How to Explain Data Preprocessing to an Interviewer

Why scale?

So a big column doesn’t dominate distance models KNN, SVM, gradient descent care

Where do you fit the scaler?

Train only Then transform test