Predictive Sales Forecasting Analysis

I have a historical sales dataset that I need turned into a reliable forecasting tool. Your task is to apply predictive analytics to uncover the underlying patterns in the data and generate accurate projections of future sales.

Here’s how I see the engagement flowing: you will start by cleaning and exploring the raw files, engineer any features that improve signal, then build and validate a forecasting model. I’m open to the techniques you feel are most appropriate—ARIMA, Prophet, gradient-boosted trees, or a custom deep-learning approach—so long as the final model proves its worth on hold-out data.

Deliverables
• A fully documented model (code + reproducible environment)
• A brief methodology report that explains assumptions, key drivers, and accuracy metrics
• Forecast outputs in both CSV format and easy-to-read visuals (Power BI, matplotlib—whichever you prefer)
• Recommendations on how I can refresh the model as new sales data comes in

Acceptance criteria: MAPE under 10 % on the most recent 3-month hold-out set, with clear instructions for rerunning the pipeline.

If this aligns with your expertise in Python, SQL, or equivalent analytics tools, I’d love to see how you would tackle the challenge and your estimated timeline.