Estimation question

How many restaurants are there in New York?

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

Fermi estimation using population and dining-frequency assumptions, adjusted for a city's known character.

State you'll estimate via NYC population and an assumed number of residents needed to support one restaurant.
Estimate NYC population at roughly 8.5 million.
Assume a ratio, e.g. roughly 1 restaurant per 400 to 500 residents based on typical urban density.
Divide population by that ratio to get an estimate.
Adjust the ratio for NYC's unusually high restaurant density and tourist-driven demand.
State the final estimate as a range, not a single overprecise figure.

Anchors on population and a stated restaurants-per-capita ratio rather than guessing the total directly.
Adjusts the ratio to reflect NYC's known unusually high restaurant density and tourist traffic.
States the assumption clearly as a range, not a single overprecise figure.
Sanity-checks against a known reference point without citing it as fact.

Using a generic per-capita ratio without adjusting for NYC's atypically high restaurant density.
Not accounting for tourist demand inflating the number of restaurants beyond what residents alone would support.

How would tourist traffic change your estimate for a city like NYC versus a suburb?
What's the biggest weakness in your assumption?