How lokavishal021/professional-sales-demand-forecasting-system works
This repository forecasts daily Online Retail sales and presents the results for operational analysis. The principal workflow is a refreshable master run: preprocessing cleans and aggregates the workbook, model training builds and saves an XGBoost model, and forecast generation produces a 30-day outlook, analytical plots, and a Power BI report. A Streamlit dashboard offers interactive analysis and can trigger a refresh. Regional aggregation and hold-out validation are distinct supporting capabilities. The dashboard excerpt is partial, so links from it to specific generated files are not asserted.
Main entry points: predict_future.py Forecast & Reporting · main.py Run Coordinator · preprocess_data.py Sales Preprocessing · sales_forecasting.py Feature Engineering & Training
Components 14
Forecast Workflow
- Run Coordinator
main.py - Sales Preprocessing
preprocess_data.py - Retail Workbook
raw_data/Online Retail.xlsx - Daily Sales Data
data/processed_daily_sales.csv - Regional Sales
data/regional_sales.csv
Modeling
- Feature Engineering & Training
sales_forecasting.py - Trained Model
models/sales_model.joblib - Backtest Results
data/validation_results.csv
Forecast Outputs
- Forecast & Reporting
predict_future.py - Power BI Report
data/powerbi_master_report.csv - Forecast Charts
Analysis Interfaces
- Streamlit Dashboard
dashboard.py - Power BI
Other components
- Retail Analyst
Connections 17
- Retail Analyst → Run Coordinator: runs pipeline
README.md - Run Coordinator → Sales Preprocessing: calls preprocessing
main.py - Run Coordinator → Feature Engineering & Training: calls training
main.py - Run Coordinator → Forecast & Reporting: calls forecast
main.py - Sales Preprocessing → Retail Workbook: reads workbook
preprocess_data.py - Sales Preprocessing → Daily Sales Data: writes daily data
preprocess_data.py - Sales Preprocessing → Regional Sales: writes regional totals
preprocess_data.py - Feature Engineering & Training → Daily Sales Data: reads daily data
sales_forecasting.py - Feature Engineering & Training → Trained Model: saves model
sales_forecasting.py - Feature Engineering & Training → Backtest Results: writes backtest
sales_forecasting.py - Forecast & Reporting → Daily Sales Data: reads daily data
predict_future.py - Forecast & Reporting → Trained Model: loads model
predict_future.py - Forecast & Reporting → Power BI Report: exports report
predict_future.py - Forecast & Reporting → Forecast Charts: saves charts
predict_future.py - Retail Analyst → Streamlit Dashboard: opens dashboard
README.md - Streamlit Dashboard → Run Coordinator: starts refresh
dashboard.py - Power BI → Power BI Report: imports report
README.md