Retail Supply Chain Analytics

A free, end-to-end supply chain analytics platform — generate your own synthetic dataset or upload your own data to visually explore it with 15+ auto-generated charts, ML models, and interactive dashboards.

---

What It Does

• Generate 50,000 synthetic supply chain records using generate_data.py
• Or upload your own CSV, Excel, TSV, or JSON file to the web dashboard
• Auto-generate 15+ chart types: bar, line, pie, scatter, bubble, radar, box plot, histogram, funnel, waterfall, correlation heatmap, and more
• Filter data with interactive slicers (categorical + numeric)
• Drill down into any category with one click
• Bookmark your favorite views
• Add calculated columns with custom formulas
• Switch themes — 6 built-in themes to choose from

---

Tech Stack

Tool	Purpose
Python 3.11+	Core language
Pandas / NumPy	Data manipulation
Matplotlib / Seaborn	Visualization
Scikit-learn	Random Forest classifier
Statsmodels	SARIMA time-series model
Flask / Chart.js	Interactive web dashboard

---

How to Run

Step 1 — Install Dependencies

pip install -r requirements.txt

Step 2 — Generate or Prepare Data

Option A — Generate synthetic data:

python generate_data.py

Creates retail_supply_chain_data.csv with 50,000 records.

Option B — Use your own dataset: Skip this step and upload your file directly to the web dashboard. It excepts multiple types of dataset. (CSV, TSV, TXT, XLSX, XLS, XLSM, JSON)

Step 3 — Run the Analysis Pipeline

python main_analysis.py

• Cleans and validates the data
• Engineers features (lead time, delays, financials, return flags)
• Generates EDA charts into outputs/
• Trains SARIMA demand forecasting model
• Trains Random Forest delay prediction classifier
• Exports clean_data_for_dashboard.csv

Step 4 — View the Static Dashboard

python dashboard.py

Generates dashboard_board.png — a 2×2 Matplotlib summary dashboard.

Step 5 — Launch the Web Dashboard

python web_dashboard.py

Open http://127.0.0.1:5050 in your browser and upload any dataset to get started.

---

Web Dashboard Features

Feature	How to Use
File Upload	Drag & drop or click the upload button
Filter Panel	Click "Filters" → check categories or set numeric ranges → "Apply Filters"
Data Table	Click "Data Table" → search, sort, paginate through raw data
Chart Switcher	Hover any chart card → click the swap icon (↺) → pick a new chart type
Chart Download	Hover any chart card → click the download icon (⇩)
Expand Chart	Hover any chart card → click the expand icon (⛶) for full-screen view
Drill Down	Click any bar/slice on a categorical chart → opens a detailed breakdown panel
Bookmarks	Set up filters/theme → click "Bookmarks" → name and save → restore anytime
Calculated Columns	Click "Calc Column" → enter name + formula → auto-reanalyse
Theme Switcher	Click the "Themes" dropdown in the header → pick a theme
Export CSV	Click "Export CSV" in the toolbar

---

Key Results

Model	Metric	Value
SARIMA Demand Forecast	MAPE	~2.97%
Random Forest Delay Classifier	Accuracy	~86%
Dataset Size	Records	50,000
EDA Visualizations	Count	10
Web Dashboard Chart Types	Count	15+

---

Troubleshooting

Issue	Solution
ModuleNotFoundError	Run pip install -r requirements.txt
DLL load failed on Windows	Use Python 3.12 or 3.13
Web dashboard port in use	PORT=8080 python web_dashboard.py
Flask server won't start	pip install flask

---

Author: Edge Shaw