My first instinct was to jump straight to car fleet costs and I almost forgot about the data processing side entirely.
Break the problem into two main components: the cost to collect new imagery (driving, equipment, processing) and the cost to process and integrate it into Google Maps. Use a bottom-up estimate based on New York's road network and Google's known Street View operations, then sanity-check with a top-down view.
Pro tip: Mention that Google already has much of the infrastructure and data, so the marginal cost is lower than a from-scratch estimate; also note that imagery is refreshed regularly, so 'lost' data might be partially recoverable from backups or other sources.
Define what 'lost street view imagery' means: is it all of New York, or just a subset? Assume we need to re-collect and re-process all street-level imagery for New York City's road network.
Calculate total road miles in New York (e.g., NYC has about 6,000 miles of roads; if including all state, more). Use this to determine the scale of data collection.
Break down into vehicle costs (cars, fuel, maintenance), equipment (cameras, LiDAR), and labor (drivers, operators). Estimate time required based on driving speed and coverage.
Include costs for stitching images, blurring faces/license plates, storing data, and integrating into Google Maps. Consider cloud computing and engineering labor.
Add up all costs and compare with known benchmarks (e.g., Google's overall Street View budget) to ensure the estimate is reasonable.
AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.