---

Dataset Description

E_Commerce.csv contains online retail transactions with features such as:

• OrderID: Unique identifier for each purchase
• CustomerID: Anonymized customer identifier
• OrderDate: Date and time of the order
• ProductCategory: Categorical label for the item purchased
• UnitPrice: Price per unit (in local currency)
• Quantity: Number of units sold
• Revenue: Total order value (UnitPrice × Quantity)

---

Technologies & Libraries

• Python 3.x
  • pandas, numpy for data manipulation
  • matplotlib, seaborn for EDA and plotting
  • scikit‑learn for feature engineering, modeling, and evaluation
• Jupyter Notebook
• Power BI for interactive dashboards

---

Key Steps & Findings

1. Data Cleaning & Preprocessing

   • Converted OrderDate to datetime, extracted day, month, and hour
   • Handled missing values via imputation or row removal
   • Dropped duplicate transactions

2. Exploratory Data Analysis

   • Distribution of Revenue, UnitPrice, and Quantity
   • Top‑selling product categories and monthly sales trends
   • Correlation heatmap to identify relationships between features

3. Feature Engineering

   • Created OrderHour, OrderMonth, and DayOfWeek from OrderDate
   • One‑hot encoded ProductCategory
   • Standardized numeric features (UnitPrice, Quantity, etc.)

4. Model Building & Evaluation

   • Trained four regressors:
     • Linear Regression
     • Decision Tree Regressor
     • Random Forest Regressor
     • K‑Nearest Neighbors Regressor
   • Evaluated using R² and RMSE

5. Hyperparameter Tuning

   • Performed GridSearchCV on Random Forest
   • Selected best parameters for n_estimators and max_depth

6. Conclusions & Recommendations

   • Key Drivers of Revenue: Unit price, quantity sold, and order timing
   • Best Model: Random Forest Regressor (R² ≈ 0.5, RMSE ≈ 25)
   • Business Insights:
     • Peak sales occur during weekends and holiday months
     • Certain product categories consistently outperform others
   • Next Steps: Incorporate customer demographics, marketing campaign data, and deploy model via Streamlit for real‑time revenue forecasting

---

How to Run

1. Clone the repository:

   git clone https://github.com/yourusername/Capstone‑ECommerce‑Project.git
   cd Capstone‑ECommerce‑Project