Turn spreadsheets into accurate AI forecasts on the fly

@Forecast Coach Use the Daily_Retail worksheet in the uploaded workbook. Use Date as ds and Units_Sold as y, including all 182 data rows. Forecast the next 14 days after the last observation with freq="D". Show the forecast widget and summarize the outlook. Report the empirical band coverage and its basis returned by the tool; do not assume a nominal coverage.

@Forecast Coach Use Stockouts_Zeros. Use Date as ds and Units_Sold as y; use Observation_Status only to interpret missingness. Keep every date. Send blank values for stockout or missing_report as null and retain both observed_zero values as numeric 0. Forecast 14 days after the final row with freq="D". Show the forecast widget and report the measured band coverage and basis.

@Forecast Coach Use Open_MWF, with Date as ds and Units_Sold as y. The stall opens only Monday, Wednesday and Friday. Keep every calendar date and send blank cells as null; preserve the genuine observed_zero as 0. Forecast the next 14 calendar days with freq="D". Show the raw plugin result and explain that predictions on closed dates are not applicable. Do not assume the model has a future opening-calendar feature. Report any missing-data or model-limit warnings.

@Forecast Coach Use Weekly_Orders, with Date as ds and Orders as y. Dates are Sunday week-ending labels exactly seven days apart. Forecast the next 8 weeks after the last row with freq="W". Do not relabel the data as daily. Show the forecast widget, describe the outlook, and report the empirical band coverage and its basis.

@Forecast Coach Use Monthly_Revenue, with Date as ds and Revenue_USD as y. Each date is the first day of the month represented by that monthly revenue. Forecast the next 6 months after the last observed month with freq="MS". The first forecast label should be 2026-09-01. Show the forecast widget and report the measured band coverage and its basis.

@Forecast Coach Use Multi_SKU and forecast only SKU_1041_Blue_Units using Date as ds. Do not use SKU_1042_Red_Units, SKU_2200_Bundle_Units, or a combined total. Forecast the next 14 days with freq="D" and show the forecast widget. Report the empirical band coverage and its basis.

@Forecast Coach Use the Cafe_Year worksheet in the uploaded workbook. Use Date as ds and Cups_Sold as y, including all 365 data rows. Forecast the next 90 days (about three months, 2026-09-18 to 2026-12-16) with freq="D", horizon=90. Show the forecast widget and summarize the quarter ahead month by month. Report the empirical band coverage and its basis returned by the tool; do not assume a nominal coverage.

@Forecast Coach Use the Holiday_Quarter worksheet in the uploaded workbook. Use Date as ds and Units_Sold as y, including all 731 data rows from 2024-09-30 to 2026-09-30, and send the four blank closed-day cells as null. Forecast the holiday quarter, 2026-10-01 to 2026-12-31, with algorithm="prophet", freq="D", horizon=92; holidays are fitted only by Prophet. From the holiday table in E1:K9, send every open row as one holidays entry {name: Holiday, dates: [Past_Date_1, Past_Date_2, Future_Date], lower_window, upper_window}, and put all three dates of every closed row in closed_dates. Show the forecast widget and summarize October, November and December, pointing out Halloween, Black Friday, the Christmas rush and the closed days. Explain in one plain sentence that Prophet learns a yearly pattern only from at least two full years of history: this sheet has two, while the one-year Cafe_Year sheet is too short for one. Report each holiday effect the tool returns, and the empirical band coverage and its basis; do not assume a nominal coverage.