Sanaesaber/morocco-sales-forecasting

Sales forecasting using SARIMA and Prophet — comparative study # Sales Forecasting: SARIMA and Prophet Comparative forecasting study using SARIMA and Facebook Prophet on e-commerce sales data with Moroccan market context. ## Business Problem How can we forecast future sales for inventory planning and budget decisions? ## Results | Model | Method | |-------|--------| | SARIMA | Classical Box-Jenkins — strong on stable trends | | Prophet | Meta model — handles seasonality and breaks | | Evaluation | MAE and RMSE benchmarked on both models | ## Tech Stack Python · Statsmodels · Facebook Prophet · Pandas · Matplotlib ## Key Insights - SARIMA performs better on linear stable trends - Prophet handles structural breaks more gracefully - Both models agree on forecast direction - Reproducible pipeline with confidence intervals ## Author Sanae Saber — Data Analyst and ML Engineer 📧 sanaesaber03@gmail.com 🔗 github.com