Estimation question

Estimate the number of tube lights used in a big city (10 Million Population).

Practice this question out loud. An AI interviewer asks it, follows up like a real interviewer would, and scores your answer. Type or speak.

What this question tests

Tests structured estimation, breaking a city-scale physical quantity into per-building and per-room assumptions.

State the funnel: population, households, buildings/units per household type, then tube lights per unit or room.
Assume average household size of 4, giving roughly 2.5 million households for a 10 million population city.
Assume an average home has 4 rooms with tube lights (living, bedrooms, kitchen), so roughly 10 million tube lights in homes alone.
Add commercial and public spaces: offices, shops, and streetlights, assumed roughly equal in count to another 20-30% of the residential figure, adding about 2.5 to 3 million more.
Total the estimate: roughly 12 to 13 million tube lights across the city, stated as an order-of-magnitude figure, not a precise count.

Shows the full breakdown (households, rooms per household, commercial add-on) rather than a single guessed number.
Labels every assumption explicitly, e.g. 'assume household size of 4', so the interviewer can challenge any specific link.
Separates residential from commercial/public tube light usage, since lumping them together would hide a major source of undercounting.
States the final answer as an order of magnitude (12-13 million) rather than false precision, appropriate for a Fermi estimate.

Jumping to a single guessed number without a visible chain of assumptions.
Ignoring commercial, office, and street lighting and counting only residential use.

How would this estimate change in a city with more LED adoption replacing tube lights?
How would you validate this estimate against real electricity or retail sales data?