Who asked: Dana Whitlock, marketing lead at Fernhill Coffee Roasters.
Why: the quarterly review decides next quarter's ad budget, and Dana wants the numbers from the export, not from the ad platforms' own dashboards.
The data: campaigns, one row per channel and month for January to April 2026: spend as text with dollar signs, clicks, orders and revenue. The February newsletter was skipped, so its row is all zeros.
What done looks like: one script that cleans the export, adds cost per order and return on ad spend without dividing by zero, compares channels on their totals, shows month-on-month change and each channel's best month, draws a faceted revenue chart in dollars, and names the channel with the best return.
This project uses the whole Advanced toolkit on one marketing export: cleaning, safe division, grouped summaries, lag(), ranks with ties, facets and dollar scales. Each step adds to the same script.