Build a predictive analytics solution to assess and minimize late delivery risks in a global supply chain network. Utilize machine learning models (Random Forest, Linear SVC, Gradient Descent), SQL, EDA, and Power BI for data analysis, visualization, and prediction.
File: DataCoSupplyChainDataset.csv
Records: 180,519 rows
Fields: 53 columns (sales, product, customer, shipping, profit, category, region, etc.)
The project is divided into three major dashboard segments:
Key KPIs (Cards):
- Total Shipments
- Sum of Sales
- Late Delivery Percentage
Visuals:
- 📈 Line Chart: Avg. Days for Shipping (Real vs. Scheduled) by Month
- 🌍 Map: Order Count by Country
- 🧮 Bar Chart: Count of Orders by Delivery Status
- 🥧 Pie Chart: Late, Early, On-Time Breakdown
- 🔍 Bar Chart: Top 10 Product Categories with Highest Delivery Risk
Business Insight: Identify bottlenecks in shipping performance, countries with frequent delays, and high-risk categories.
Key KPIs (Cards):
- Total Unique Customers
- Average Order Value
- Customer Profitability
- Sum of Sales
Visuals:
- 📈 Line Chart: Sales per Customer Over Time
- 🛍️ Bar Chart: Top Products by Sales Count
- 🎯 Bar Chart: Top Products by Revenue
- 📊 Clustered Column: Product Category vs Region
- 🗂️ Filters: Year, Department Name
Business Insight:
Track loyal customers, most profitable product lines, customer purchase frequency, and geographical demand patterns.
Key KPIs (Cards):
- Sum of Sales
- Count of Sales
Visuals:
- 📈 Line Chart: Sales & Benefit per Order by Month
- 🚚 Bar Chart: Shipping Mode by Delivery Status
- 🌍 Map: Average Sales by Market
- 🏷️ Column Chart: Sales by Customer Segment
- 🥧 Pie Chart: Shipping Mode by Payment Type
Business Insight: Monitor top-selling items, customer segment contribution, shipping performance, and payment method popularity.
- Year
- Delivery Status
- Department Name
- Country/Region
- Customer Segment
- Removed sensitive columns (e.g., emails, passwords)
- Converted date fields to datetime format
- Handled null values in product and location columns
- Created new calculated columns (e.g., Delivery Delay, AOV)
- Top revenue-generating regions and products
- Delivery delay patterns by geography and category
- Loyal vs. one-time customers
- Preferred shipping modes and payment types