ResilioChain 🔗

End-to-End Supply Chain Intelligence Platform

"Built to optimize Service Levels while minimizing Holding Costs."

---

📌 Problem Statement

Modern businesses lose billions due to Stock-outs (shelves going empty) and Dead Stock (capital frozen in unsold inventory). ResilioChain solves both by building a Digital Twin of a warehouse — predicting future demand and automating ordering decisions.

---

🏗️ Project Structure

ResilioChain/
├── src/
│   ├── data_generator.py       # Phase 1 — Synthetic 365-day warehouse simulation
│   ├── eda.py                  # Phase 2 — EDA, ABC Analysis, Lag Features (upcoming)
│   ├── forecast.py             # Phase 3 — XGBoost demand forecasting (upcoming)
│   ├── optimizer.py            # Phase 4 — ROP / Safety Stock engine (upcoming)
│   └── app.py                  # Phase 5 — Streamlit dashboard (upcoming)
├── data/
│   ├── transactions.csv        # 1,825 rows — daily sales per product (cleaned)
│   ├── inventory.csv           # 1,825 rows — daily stock levels per product
│   ├── products.csv            # 5 rows — product reference table
│   └── suppliers.csv           # 3 rows — supplier reference table  # Transactions & Products EDA   # Inventory & Suppliers EDA
├── notebooks/
│   └── eda/
│       ├── session_2_eda_loading_and_inspection_part_A.ipynb        # Transactions & Products — loading, types, nulls, drops
│       ├── session_2_eda_loading_and_inspection_part_B.ipynb        # Inventory & Suppliers — loading, types, nulls, drops
│       ├── session_2_eda_grouping_and_aggregations_part_A.ipynb     # Transactions & Products — groupby, aggregations, trends
│       └── session_2_eda_grouping_and_aggregations_part_B.ipynb     # Inventory & Suppliers — stockout rates, supplier performance
├── outputs/                    # Charts and analysis (generated)
├── requirements.txt
└── README.md

---

🚀 Project Phases

Phase	Focus	Status
1 — Data Engine	Synthetic 365-day simulation	✅ Complete
2 — EDA	Data Loading, Inspection & Quality Audit	✅ Complete (Sessions A & B)
3 — Forecasting	XGBoost demand prediction (30-day)	⏳ Upcoming
4 — Optimization	ROP, Safety Stock, EOQ	⏳ Upcoming
5 — Deployment	Streamlit + What-If Dashboard	⏳ Upcoming

---

✅ What's Been Built (Sessions 1 & 2)

Session 1 — Data Engine (src/data_generator.py)

Generates a fully synthetic but realistic warehouse simulation across 365 days for 5 products and 3 suppliers.

Product Catalogue:

ID	Product	Category	Unit Cost	Unit Price	Base Demand/day
P001	Wireless Headphones	Electronics	$45.00	$89.99	22
P002	Yoga Mat	Fitness	$12.00	$29.99	18
P003	Stainless Water Bottle	Kitchen	$8.00	$19.99	30
P004	Bluetooth Speaker	Electronics	$30.00	$59.99	15
P005	Winter Jacket	Apparel	$55.00	$129.99	10

Supplier Table:

ID	Supplier	Lead Time	Reliability
S001	AsiaTech Imports	14 days	92%
S002	EuroGoods Ltd	7 days	97%
S003	LocalFast Supply Co	3 days	99%

Key simulation features:

• Poisson-distributed demand (realistic "noisy" sales, not flat)
• Day-of-week seasonality (30% weekend boost)
• Monthly seasonality (Dec spike ×1.60, Jan slump ×0.75, summer rise)
• Running inventory simulation with reorder point (ROP = 150 units), order qty = 400
• Supplier lead time noise — delayed shipments modelled stochastically
• Generates 4 CSV files: transactions.csv, inventory.csv, products.csv, suppliers.csv

---

Session 2 — EDA: Data Loading & Inspection (Parts A & B)

Part A — notebooks/eda/session_2_eda_loading_and_inspection_part_A.ipynb

Deep inspection of transactions.csv and products.csv:

• Shape validation: transactions.csv → 1,825 rows × 9 columns (5 products × 365 days)
• Zero null values, zero duplicates — dataset confirmed clean
• Date range confirmed: 2024-01-01 → 2024-12-31
• Demand statistics: Min 4 | Max 72 | Mean 22.33 | Std Dev 10.82 — significant daily fluctuation
• Redundant columns dropped: product_name, category, unit_cost, unit_price (static reference data better kept in products.csv)
• Mathematical consistency verified: gross_profit = revenue − cogs holds for all rows
• Profit margin analysis per product (Yoga Mat leads at ~60%)
• transactions.csv re-saved in cleaned, lean format

Part B — notebooks/eda/session_2_eda_loading_and_inspection_part_B.ipynb

Deep inspection of inventory.csv and suppliers.csv:

• Shape validation: inventory.csv → 1,825 rows × 7 columns (daily diary: one row per product per day)
• suppliers.csv → 3 rows × 4 columns — confirmed static reference table (not a time-series)
• Column semantics documented: closing_stock, stockout_flag, reorder_triggered, lead_time_days
• Confirmed the data structure underpinning ROP simulation

---

🗂️ Dataset Schema

transactions.csv (cleaned)

Column	Type	Description
date	object → datetime	Transaction date
product_id	object	Product identifier (P001–P005)
units_sold	int	Daily units sold (Poisson-distributed)
revenue	float	units_sold × unit_price
cogs	float	units_sold × unit_cost
gross_profit	float	revenue − cogs

inventory.csv

Column	Type	Description
date	object	Date of snapshot
product_id	object	Product identifier
closing_stock	int	End-of-day stock level
stockout_flag	int	1 if stock = 0, else 0
reorder_triggered	int	1 if reorder fired on this day
supplier_id	object	Assigned supplier
lead_time_days	int	Base lead time (before noise)

---

🛠️ Tech Stack

Layer	Tools
Data Engineering	Python, Pandas, NumPy, Faker
Visualization	Matplotlib, Seaborn, Plotly
AI / Forecasting	XGBoost, Scikit-learn
Optimization	SciPy
Deployment	Streamlit
Notebooks	JupyterLab
DevOps	Git, GitHub

---

▶️ Quick Start

git clone https://github.com/dominator959/ResilioChain.git
cd ResilioChain

# Create and activate environment
conda create -n resiliochain python=3.11 -y
conda activate resiliochain

# Install dependencies
pip install -r requirements.txt

# Phase 1 — Generate all datasets
python src/data_generator.py

# Phase 2 — Open EDA notebooks
jupyter lab notebooks/eda/

---

📊 Commit History

Date	Session	Description	Author
2026-06-01	2	Complete Session 2 EDA — Part A (transactions + products inspection)	faizantoheed456
2026-06-01	2	Complete Session 2 EDA — Part B (inventory + suppliers inspection)	dominator959
2026-05-31	1	Session 1: Generating a Data Engine (data reformatted)	faizantoheed456
2026-05-31	1	Session 1: Generating a Data Engine	faizantoheed456
2026-05-31	—	Revise project structure formatting in README	dominator959
2026-05-31	—	Add project folder structure	dominator959
2026-05-31	—	Add professional README with project roadmap	dominator959
2026-05-31	—	Initial project structure and requirements	dominator959
2026-05-31	—	Initial commit	dominator959

---

💼 Business Value

1. Capital Efficiency — Frees frozen cash from over-stocked warehouses
2. Customer Loyalty — Eliminates stock-outs so shelves are always full
3. Operational Clarity — CEO dashboard shows the future, not just the past

---

👥 Team

Name	GitHub
Muhammad Usman	@dominator959
Faizan Toheed	@faizantoheed456

---

Supply Chain Data Science Portfolio Project — actively being developed.