Separating Weather From Other Drivers of Consumer Spending Across a Winter Season

You are asked to work out how much the weather affected consumer spending over the past winter. Explain how you would scope the problem, which data you would use, and how you would separate the weather's effect from everything else that moves spending. ```hint Decide what you compare against An effect only exists against a baseline. Before choosing data, decide what spending would have looked like had the weather been ordinary for that place and time of year, and how you could observe that. ``` ### Clarifying Questions - Which spending: all consumer spending across retail and services, one category, or only the spending that flows through one company's products? - Which geography: one country, or several regions with very different winters? - What does "impact" mean here: the difference from a normal winter, the gap against a forecast, or the effect of individual events such as storms and cold snaps? - What decision will the result feed, such as forecasting, inventory planning or marketing budgets? ### What a Strong Answer Covers - A scoped definition of spend, geography, period and "impact", with the baseline it is measured against - Spending and weather data at matching levels of detail, and the gaps in their coverage - A way to separate weather from seasonality, holidays, economic conditions and other confounders - The difference between spending that was shifted in time or channel and spending that was lost or created - How the estimate is validated, what uncertainty it carries, and how it would be used ### Follow-up Questions - How would you tell whether spending lost during a storm week came back later? - How would your approach change if you had only one company's transaction data rather than economy-wide figures? - How would you use the result to plan for next winter?