I started with US population, tried to narrow down to people likely to go out on Halloween, then layered in what fraction might use Lyft vs drive or walk.
Break the problem into supply and demand sides: estimate the number of active Lyft drivers on Halloween and the average number of rides each driver completes, then multiply. Alternatively, estimate from the rider side by segmenting the population and applying adoption and ride frequency assumptions. Clearly state your assumptions and validate with a sanity check against known data (e.g., total Lyft rides per day).
Pro tip: Mention that Halloween demand is highly concentrated in the evening hours (e.g., 6 PM–2 AM) and that ride frequency per driver may be higher than average due to surge pricing and increased demand, but also note that many drivers may choose not to work that night, affecting supply.
Define the geographic scope (e.g., US only or global) and time frame (the entire day or just peak hours). State that you'll estimate for a typical Halloween in a mature market like the US.
Decide whether to estimate from the supply side (drivers) or demand side (riders). For a quick estimate, the supply side is often simpler: number of active drivers × average rides per driver.
Start with the total number of Lyft drivers in the US, then estimate the percentage who are active on Halloween night. Consider that some drivers may avoid the night due to safety or traffic, while others may be attracted by higher earnings.
Determine the average number of rides a driver completes during a typical shift, then adjust for Halloween-specific factors like higher demand, shorter trips, and surge pricing. For example, if a driver normally does 2 rides per hour, on Halloween they might do 3 due to increased demand.
Multiply the number of active drivers by the average rides per driver to get the total. Compare this to known benchmarks (e.g., Lyft's total rides per day) 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.