A Data Analytics and Machine Learning project that analyzes logistics delivery data to identify the major factors affecting delivery costs and provides business insights for cost optimization.
This project performs Exploratory Data Analysis (EDA), Feature Engineering, Machine Learning, and Power BI Dashboard Visualization on an Indian logistics delivery dataset.
The objective is to understand delivery operations, identify the key drivers of delivery cost, compare multiple machine learning models, visualize operational performance through an interactive dashboard, and generate actionable business recommendations.
- Analyze logistics delivery data
- Perform data cleaning and preprocessing
- Explore delivery patterns through visualizations
- Engineer features for machine learning
- Train and compare regression models
- Build an interactive Power BI dashboard
- Identify key delivery cost drivers
- Generate business insights and recommendations
Source: Kaggle
Dataset: Indian Multi-Partner Delivery Logistics Dataset
The dataset includes:
- Delivery Partner
- Package Type
- Vehicle Type
- Delivery Mode
- Region
- Weather Condition
- Distance (km)
- Package Weight (kg)
- Delivery Rating
- Delivery Status
- Delivery Cost
- Python
- Pandas
- NumPy
- Matplotlib
- Seaborn
- Scikit-learn
- Google Colab
- Microsoft Power BI
Dataset Collection
│
▼
Data Cleaning
│
▼
Exploratory Data Analysis
│
▼
Feature Engineering
│
▼
Data Preprocessing
│
▼
Machine Learning Models
│
▼
Model Evaluation
│
▼
Power BI Dashboard
│
▼
Business Insights
│
▼
Business Recommendations
The following models were trained and evaluated:
- Linear Regression
- Decision Tree Regressor
- Random Forest Regressor
- Mean Absolute Error (MAE)
- Root Mean Squared Error (RMSE)
- Median Prediction Error
An interactive Power BI dashboard was developed to visualize logistics performance and delivery operations.
The dashboard includes:
- KPI Cards
- Total Deliveries
- Average Delivery Cost
- Average Delivery Distance
- Average Delivery Rating
- Average Delivery Cost by Region
- Average Delivery Cost by Vehicle Type
- Delivery Status Distribution
- Package Type Distribution
- Delivery Partner Performance
- Weather-wise Delivery Cost Analysis
- Interactive Filtering using slicers
Logistics_Cost_Analytics_Dashboard.pbix
- Delivery distance is the strongest factor influencing delivery cost.
- Package weight significantly affects transportation expenses.
- Delivery mode contributes to logistics cost variations.
- Random Forest achieved the lowest prediction error among the evaluated models.
- The Power BI dashboard provides interactive insights into delivery performance and operational efficiency.
- Optimize long-distance delivery routes.
- Improve vehicle allocation based on package weight.
- Optimize delivery mode selection.
- Monitor delivery performance using interactive dashboards.
- Use historical logistics data for delivery cost estimation.
- Support operational planning through predictive analytics.
The complete notebook, including all visualizations, machine learning models, outputs, and business insights, can be viewed here:
Logistics_Cost_Analytics/
├── Logistics_Cost_Analytics.ipynb
├── Logistics_Cost_Analytics_Dashboard.pbix
├── dashboard.png
├── README.md
└── requirements.txt
- Route Optimization
- Real-Time Cost Prediction
- Weather API Integration
- Fuel Price Analysis
- Interactive Dashboard Enhancement
- Deployment as a Web Application
Reva Digraskar
B.Tech, Electronics and Communication Engineering
National Institute of Technology Jamshedpur
GitHub: https://github.com/ugec67