Case study
Revenue Forecasting & Seasonality
Time-series forecast of monthly revenue with Prophet, decomposing trend and seasonality — visualized so non-technical stakeholders could plan headcount and inventory with confidence.
What's the fastest way to let anyone explore revenue & seasonality for themselves?
Using Python, Prophet, Pandas, and Plotly, the data was transformed into a visual narrative — time-series forecast of monthly revenue with Prophet, decomposing trend and seasonality — visualized so non-technical stakeholders could plan…
The full solution — including source code and documentation — is available on GitHub.
Hover a row to edit · changes save to your portfolio
Goal
Give the ops team a defensible 6-month revenue forecast instead of last-year-plus-10%.
Approach
- Cleaned 3 years of daily sales, handled holidays and promos as regressors
- Fit Prophet, tuned changepoint prior via rolling backtests
- Built an interactive Plotly view of trend, weekly, and yearly seasonality
Result
Backtested MAPE of 7.8% on held-out months, and the seasonality breakdown finally explained the recurring Q1 dip stakeholders kept re-litigating.