City Bike Stations | Operations Dashboard

simulated-bike-ops-dashboard-2026-09.xlsx — an editable, fully formula-driven workbook built on 42 simulated ride records (14 days × 3 fictional stations, 1–14 Sep 2026, fixed random seed 20260901).

Sheets (every one carries the banner "Simulated data · Not real operations" in row 1, plus the same text in the print header):

• Read Me & Method — purpose, provenance, sheet index, definitions and units, full formula map, assumptions, how to edit, known limits.
• Daily Records — the 42 rows. Amber cells are editable inputs (Date, Day, Station, Station ID, Bikes in Service, Rides, Avg Fare, Avg Duration, Avg Distance); grey cells are formulas: Revenue = Rides × Avg Fare, Revenue per Bike = Revenue / Bikes in Service, Orders per Bike = Rides / Bikes, Revenue per Order = Revenue / Rides. Row 48 totals with SUM; whole-number validation on Bikes and Rides; frozen header, autofilter, repeated print titles.
• Station Summary — SUMIFS for total rides and revenue, COUNTIFS for days reported, AVERAGEIFS for average daily rides, duration and distance, plus revenue per bike, orders per bike, revenue per order and average daily revenue. Fleet size is an editable input.
• Trends & Charts — a SUMIFS chart-data block (per-station revenue and total rides by day) feeding four charts with titles, axis titles and units: daily revenue by station (line), daily total rides (column), revenue per bike (bar), orders per bike (bar).
• QA Checks — 12 in-workbook checks (record count, distinct stations/dates, revenue identity per row, summary-to-records ties, revenue/orders per bike identities, chart-data tie, error-token scans), each returning PASS/FAIL plus an overall verdict.

Simulated results (14 days): 4,655 rides, $17,223.58 revenue. Riverside Dock (42 bikes) 2,171 rides / $8,422.91 / $200.55 per bike / 51.69 orders per bike; Maple Gate (30 bikes) 1,533 / $5,626.69 / $187.56 / 51.10; Old Mill Yard (24 bikes) 951 / $3,173.98 / $132.25 / 39.63. Weighted average fare $3.70 per order.

Verification on the final saved bytes: package opens clean, 5 sheets, 330 formula cells, 4 charts, no error tokens, all 12 QA checks PASS. Every formula cache was recomputed independently in Python from the input columns and matches to 1e-9 (so values show correctly before recalculation; the workbook also recalcs on open). Column widths are explicit on every used column and a deterministic text-fit audit reports zero clipped or overflowing labels.

Coverage limit: no LibreOffice/renderer exists in this Computer, so I could not produce a rendered page preview — visual QA was a deterministic width/content audit rather than a rendered inspection. If you open it and want spacing adjusted, tell me which sheet.

Also: the data is invented; the workbook is a demonstration template, not a real-operations or forecasting tool.