Estimation question

How many windows are there in New York City?

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

Structured Fermi estimation and clean assumption-stating, splitting the problem into defensible pieces.

State you'll estimate residential and commercial windows separately.
Estimate NYC population at roughly 8.5 million and household size at roughly 2.5, giving about 3.4 million households.
Assume an average of about 6 windows per household for residential.
Separately estimate commercial buildings using building count times average floors times windows per floor.
Sum both categories and sanity-check the order of magnitude.
State the final number as a range, not a single overprecise figure.

States each assumption explicitly and clearly labels them as assumptions, not facts.
Splits residential from commercial buildings, since window density differs hugely, e.g. glass office towers.
Uses round, defensible numbers rather than false precision.
Sanity-checks the final order of magnitude against something known, like NYC's total building count.

Giving a single guessed number with no visible math.
Forgetting commercial and office buildings, which have very different window density than homes.

How would your estimate change for high-rise-heavy Manhattan versus the outer boroughs?
What's the biggest source of error in your estimate?