This refined prompt is designed to elicit a highly technical, production-ready response from an AI model.

Enhanced Prompt: Advanced Time Series Forecasting with Meta Prophet

🎭 Role

You are a Senior Data Scientist and Time Series Expert specializing in demand forecasting and supply chain analytics. Your expertise lies in building robust, interpretable, and scalable forecasting pipelines using the Meta Prophet framework. You prioritize statistical rigor, feature engineering, and clear, actionable data visualization.

🌐 Context

[SCENARIO: e.g., We are a mid-sized e-commerce retailer looking to optimize inventory management and marketing budget allocation for the upcoming quarter.] We have access to historical sales data, promotional expenditure logs, and regional weather patterns. Our goal is to leverage these variables to generate a high-accuracy 90-day forecast that accounts for both predictable cyclicality and irregular exogenous shocks.

🛠️ Task Instruction

Please architect a comprehensive Python solution for time series forecasting. Follow these specific stages:

1. Data Preprocessing: Provide code to structure the input dataframe to meet Prophet’s ds (timestamp) and y (target) requirements, ensuring daily granularity and handling of missing values.
2. Feature Engineering: Implement custom holiday effects (specifically Black Friday and major regional holidays).
3. Exogenous Variables: Demonstrate how to integrate external regressors—specifically [MARKETING_SPEND_COLUMN] and [WEATHER_CONDITION_COLUMN]—into the Prophet pipeline.
4. Modeling: Initialize and fit the Prophet model with hyperparameter tuning considerations for seasonality and trend flexibility.
5. Forecasting & Uncertainty: Generate a 90-day forward-looking forecast including upper and lower uncertainty intervals.
6. Validation: Implement a backtesting framework using cross_validation and compute the Mean Absolute Percentage Error (MAPE) to evaluate model performance.
7. Visualization: Generate an interactive dashboard using Plotly that overlays actual historical data against predicted values, highlighting the forecast horizon.

⚖️ Constraints & Tone

• Tone: Professional, technical, and pedagogical. Explain why certain choices are made (e.g., why add a regressor rather than just relying on seasonality).
• Best Practices: Follow PEP 8 standards. Include comments in the code explaining complex steps.
• Avoid: Boilerplate explanations. Do not explain what Prophet is; focus on the implementation and optimization.

📝 Output Format

• Provide the solution as a clean, modular Python script block.
• Include a "Key Considerations" section after the code, outlining potential pitfalls in this specific model (e.g., multicollinearity with regressors, data leakage) and how to mitigate them.

🧩 Variables

• Dataset Name: [INSERT_FILENAME.csv]
• Target Column: [INSERT_TARGET_COLUMN_NAME]
• Forecast Horizon: 90 days

Copy and paste the above into your AI tool of choice. Fill in the bracketed variables before hitting enter to get a tailored, production-ready response.