Income Eats the Vote

You're clustering customers using two raw features: annual income in dollars (range: tens of thousands) and age in years (range: 0-100). When you compute plain Euclidean distance between customer pairs, a $50,000 income gap swamps a 40-year age gap in the total distance, even though age should matter too. What's the best fix before applying Euclidean distance?

Take the log of income only and leave age as-is

Standardize (or min-max scale) both features so each contributes comparably to the distance

Switch to a decision tree model instead, since it sidesteps the distance calculation entirely

Replace Euclidean distance with L-infinity distance, which will automatically rebalance the features