Duration: 10 weeks
Target Audience: Supply chain professionals, data analysts, and data scientists looking to apply time series forecasting and create interactive dashboards for demand planning and inventory management.
Course Description: This course teaches participants how to forecast demand in the supply chain industry using time series models and build interactive Shiny dashboards in R to visualize and share insights. Through hands-on projects, learners will master data preparation, forecasting techniques, and dashboard creation, culminating in a capstone project that integrates forecasting with a deployable dashboard.
Curriculum
- Week 1: Introduction to Supply Chain Analytics and Demand Forecasting
- Overview of supply chain analytics: inventory management, demand planning, and logistics.
- Importance of demand forecasting in optimizing supply chains.
- Introduction to R and required packages (forecast, shiny, dplyr, ggplot2).
- Assignment: Set up R environment and explore a sample supply chain dataset (e.g., sales data).
- Week 2: Data Preparation for Forecasting
- Importing and cleaning supply chain data (e.g., sales, shipments, inventory levels).
- Handling missing values, outliers, and seasonality in time series data.
- Tools: R (dplyr, tidyr, lubridate for date handling).
- Assignment: Clean a dataset of historical sales data and identify trends.
- Week 3: Fundamentals of Time Series Analysis
- Components of time series: trend, seasonality, and noise.
- Decomposition of time series data using R (stl, decompose).
- Visualizing time series data with ggplot2.
- Assignment: Decompose a time series dataset and interpret its components.
- Week 4: Basic Forecasting Models
- Introduction to simple forecasting methods: moving averages, exponential smoothing.
- Implementing these models in R using the forecast package (ses, holt).
- Evaluating forecast accuracy: MAE, RMSE, MAPE.
- Assignment: Apply exponential smoothing to a sales dataset and compare accuracy metrics.
- Week 5: Advanced Time Series Models – ARIMA
- Understanding ARIMA models: Auto-Regressive, Integrated, Moving Average components.
- Stationarity and differencing (adf.test, diff in R).
- Building and fitting ARIMA models (auto.arima in forecast package).
- Assignment: Fit an ARIMA model to a supply chain dataset and forecast demand for the next 12 months.
- Week 6: Seasonal Models and Advanced Techniques
- Seasonal ARIMA (SARIMA) for handling seasonality in demand data.
- Introduction to Prophet for forecasting (prophet package in R).
- Cross-validation for time series (tsCV in forecast package).
- Assignment: Use SARIMA or Prophet to forecast seasonal demand and evaluate performance.
- Week 7: Introduction to Shiny Dashboards
- Basics of Shiny: structure of a Shiny app (UI, server).
- Building a simple dashboard to display time series data.
- Tools: R (shiny, shinydashboard).
- Assignment: Create a Shiny app to visualize raw sales data with interactive filters (e.g., date range).
- Week 8: Enhancing Dashboards with Forecasting Insights
- Integrating forecasting results into a Shiny dashboard.
- Adding interactive elements: dropdowns for selecting models, sliders for forecast horizons.
- Visualizing forecasts with plotly for interactivity.
- Assignment: Build a dashboard that displays both historical data and ARIMA/Prophet forecasts.
- Week 9: Deployment and Best Practices
- Deploying Shiny apps for stakeholder access (shinyapps.io or local hosting).
- Best practices for dashboard design: clarity, interactivity, and performance.
- Handling large datasets in Shiny (e.g., data aggregation, reactive expressions).
- Assignment: Deploy a Shiny app and share the link for feedback
- Week 10: Capstone Project
- Project: Build a demand forecasting system for a supply chain scenario (e.g., retail inventory).
- Steps: Clean the data, apply time series models (ARIMA or Prophet), evaluate forecasts, and create a Shiny dashboard to present results.
- Deliverable: A deployed Shiny app with documentation and a presentation summarizing the forecasting model and insights.
- Final review and feedback session.
