Supply Chain Optimization: Demand Forecasting Application

Overview

This application empowers supply chain professionals to optimize inventory management and demand forecasting using advanced data-driven analytics. Leveraging historical sales data from Rossmann stores and market sector performance provides actionable insights to enhance supply chain efficiency and decision-making.

Purpose

Supply chain managers consistently face challenges in balancing inventory levels:

• Excess inventory ties up valuable capital.
• Insufficient inventory results in stockouts and lost sales.

This application addresses these challenges by:

• Analyzing historical sales data to forecast future demand.
• Providing inventory recommendations based on lead times and reorder thresholds.
• Integrating market performance indicators to understand external impacts on demand.
• Supporting informed, data-driven decision-making to optimize supply chain performance.

Features

• Data Import & Processing: Seamlessly import and preprocess Rossmann store sales data.
• Demand Forecasting: Accurately forecast future product demand using machine learning.
• Market Analysis: Incorporate external market data (Alpha Vantage API) to identify trends affecting demand.
• Inventory Recommendations: Calculate optimal inventory levels based on forecasted demand and configurable parameters.
• Interactive Visualization: Visualize trends, forecasts, and inventory levels using intuitive charts and tables.
• Customizable Settings: Adjust lead times, reorder thresholds, forecast horizons, and machine learning parameters easily.

Technologies

• Frontend: WPF (Windows Presentation Foundation), XAML
• Backend: C#, .NET Framework
• Machine Learning: ML.NET (Regression, Forecasting Algorithms), optional LSTM via external integration
• Data Handling: CSV file processing, structured data models
• External API: Alpha Vantage (Market Data Integration)

Mathematical Equations

Economic Order Quantity (EOQ)

Determines optimal order quantity minimizing total inventory costs:

$$ EOQ = \sqrt{\frac{2DS}{H}} $$

• ( D ): Annual demand
• ( S ): Ordering cost per order
• ( H ): Holding cost per unit per year

Reorder Point (ROP)

Indicates inventory level to trigger a new order:

$$ ROP = (d \times L) + SS $$

• ( d ): Average daily demand
• ( L ): Lead time (in days)
• ( SS ): Safety stock

Safety Stock (SS)

Buffer stock to handle variability in demand and supply:

$$ SS = Z \times \sigma_{dL} $$

• ( Z ): Desired service level (number of standard deviations, e.g., Z = 1.65 for 95%)
• ( \sigma_{dL} ): Standard deviation of demand during lead time

Forecasting Accuracy Metrics

Root Mean Squared Error (RMSE)

$$ RMSE = \sqrt{\frac{1}{n}\sum_{i=1}^{n}(y_i - \hat{y_i})^2} $$

Mean Absolute Error (MAE)

$$ MAE = \frac{1}{n}\sum_{i=1}^{n}|y_i - \hat{y_i}| $$

• ( y_i ): Actual value
• ( \hat{y_i} ): Forecasted value
• ( n ): Number of observations

Installation

Prerequisites:

• Windows OS
• .NET Framework 4.7.2 or later
• Visual Studio 2019 or later (recommended for development)

Steps:

1. Download or clone the repository.
2. Extract contents to your preferred location.
3. Ensure the data files (train.csv, test.csv, store.csv) are placed in the Data folder.
4. Open the solution (.sln) file with Visual Studio, or execute DemandForecastingApp.exe directly.

How to Use

Step 1: Configure Application

• Go to the Settings tab.
• Enter your Alpha Vantage API key (obtain your own for higher reliability).
• Set your desired forecast horizon (default: 12 months).
• Adjust ML model parameters as required.
• Click Save Settings.

Step 2: Load Data

• Click Load Data.
• The application will import:
  • train.csv: Historical sales data.
  • test.csv: Validation dataset.
  • store.csv: Store metadata.
• Verify successful data import via the confirmation message.

Step 3: Run Forecast

• Enter your desired Lead Time and Reorder Threshold.
• Click Run Forecast.
• Wait for processing; the application generates demand forecasts and inventory recommendations.

Step 4: Analyze Results

• Navigate between tabs:
  • Forecast Details: View detailed forecasting results.
  • Inventory Recommendations: Evaluate suggested inventory levels.
  • Market Analysis: Review sector performance and its impact on demand.
• Use visualizations to identify patterns and insights.
• Refresh market data using the Refresh Data button as necessary.

Data Files

Place the following files in the Data folder:

• train.csv: Historical sales data (columns: Store, DayOfWeek, Date, Sales, Customers, Open, etc.).
• test.csv: Validation dataset (structure similar to train.csv).
• store.csv: Store metadata (StoreType, Competition details, Promo information, etc.).

Troubleshooting

• Data Loading Issues: Verify CSV formatting and file placement in the Data folder.
• API Key Issues: Confirm your Alpha Vantage API key in the Settings tab.
• Forecasting Errors: Ensure correct data import before executing forecasts.
• UI Issues: Adjust theme or display settings for improved visibility.

Developer Notes

• Main Logic: Located in MainViewModel.cs (MVVM pattern).
• Data Importing: Managed by the RossmannDataImporter class (/Data).
• Forecast Algorithms: Contained in the ForecastEngine namespace (/Models).
• UI Architecture: Adheres to MVVM for clear separation of concerns and maintainability.

Future Enhancements

• Additional forecasting models (e.g., advanced deep learning LSTM integration)
• Reporting and data export capabilities
• ERP integration for broader system connectivity
• Mobile app companion for real-time monitoring