Estimation question

Google Streetview data for San Francisco is lost. What's the cost to replace it?

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

Estimation skill through operational cost decomposition, not the exact dollar figure.

Restate the question: cost to re-capture Street View imagery for San Francisco specifically, covering all mapped streets.
Estimate the total road mileage in San Francisco, stating a rough assumption based on the city's size and street density.
Estimate the operational cost per mile of driving and capturing imagery, including vehicle operation, camera equipment amortization, and data processing.
Multiply road mileage by cost per mile to reach a total capture cost, showing the arithmetic clearly.
Add a data processing and storage cost estimate, since raw imagery needs stitching, quality checks, and hosting.
State a final illustrative range and flag that operational cost per mile is the biggest source of uncertainty in the estimate.

Breaks the estimate into road mileage and cost per mile explicitly, rather than guessing a single lump cost figure.
States a clear per-mile cost assumption covering vehicle, equipment, and labor, acknowledging it as illustrative.
Adds a separate line for data processing and storage cost, recognizing capture is not the only cost involved.
Shows the full arithmetic chain so the reasoning can be checked step by step.

Guessing a single total cost with no mileage or per-mile cost decomposition.
Ignoring data processing and storage costs, focusing only on the physical driving and capture.
Not showing the arithmetic connecting assumptions to the final number.

How would this cost change if the outage also required updating indoor Street View imagery?
What's the biggest lever that could change this estimate significantly?