Diamond Price Prediction Project

Project Overview

💎 This project aims to predict the prices of diamonds based on their features using machine learning models. The project demonstrates a complete machine learning pipeline, including data ingestion, preprocessing, transformation, and model training.

Dataset Details

The dataset is sourced from Kaggle's Gem Price Prediction. It contains the following features:

• Categorical Features: cut, color, clarity
• Numerical Features: carat, depth, table, x, y, z
• Target Feature: price

The dataset was split into training (70%) and testing (30%) datasets.

Artifacts

📂 The following artifacts are generated and used in the project:

• raw.csv: Raw dataset.
• train.csv and test.csv: Training and testing datasets.
• preprocessor.pkl: Preprocessing pipeline for data transformation.
• model.pkl: Trained machine learning model.

Technologies Used

• Programming Language: Python 🔰
• Libraries:
  • Data Manipulation: pandas, numpy
  • Preprocessing: scikit-learn
  • Logging and Exception Handling: logging, sys

Project Workflow

1. Data Ingestion 📥

• Reads the raw dataset (gemstone.csv).
• Saves raw, training, and testing data into artifacts/ directory.
• Splits the dataset into training and testing sets with a 70-30 split.

2. Data Transformation 🔄

• Handles missing values using SimpleImputer.
• Scales numerical features using StandardScaler.
• Encodes categorical features (cut, color, clarity) using OrdinalEncoder with predefined category rankings.
• Combines pipelines into a ColumnTransformer for seamless preprocessing.
• Saves the preprocessor object as preprocessor.pkl.

3. Model Training 🤖

• Models Used:
  • Linear Regression
  • Lasso Regression
  • Ridge Regression
  • ElasticNet
  • Decision Tree Regressor
• Evaluates models using R² score.
• Selects the best model based on the highest R² score.
• Saves the best model as model.pkl.

4. Prediction Pipeline 🧪

The prediction pipeline (PredictPipeline) includes:

• Loading the saved preprocessor and model objects.
• Scaling input features and making predictions.
• Handles both manual input and data from a DataFrame.

Results 📊

• The best-performing model was identified based on R² score.
• The R² scores for all models are logged and reported.

Installation Instructions ⚙️

1. Clone the repository:

   git clone https://github.com/AryanDhanuka10/Diamond_Price_Prediction/tree/main

2. Navigate to the project directory:

   cd diamond-price-prediction

3. Install required dependencies:

   pip install -r requirements.txt

Usage 🚀

1. Run the data ingestion module to generate train and test datasets:

   python src/data_ingestion.py

2. Run the data transformation module to preprocess data:

   python src/data_transformation.py

3. Train models and select the best model:

   python src/model_trainer.py

4. Use the saved model (model.pkl) for predictions:

   from src.prediction_training import PredictPipeline, CustomData
   
   custom_data = CustomData(carat=0.5, depth=61.5, table=55, x=4.5, y=4.6, z=2.9, cut="Ideal", color="E", clarity="VVS1")
   df = custom_data.get_data_as_dataframe()
   
   predict_pipeline = PredictPipeline()
   prediction = predict_pipeline.predict(df)
   print("Predicted Price:", prediction)

Challenges and Learnings 🧪

• Challenge: Handling categorical data encoding with ordinal features.
  • Solution: Used OrdinalEncoder with custom category rankings for categorical features.
• Challenge: Ensuring robust preprocessing for both training and testing datasets.
  • Solution: Implemented a unified ColumnTransformer pipeline.

Streamlit App 🎮

Access the web application for real-time price prediction: Diamond Price Prediction App

Contribution Steps 📚

We welcome contributions to improve this project! Follow these steps:

1. Fork the repository.
2. Create a new branch:

   git checkout -b feature-name

3. Make your changes and commit them:

   git commit -m "Add detailed description of your changes"

4. Push to your branch:

   git push origin feature-name

5. Create a pull request describing your changes.

Future Scope 🌟

• Extend the project to include additional models such as Random Forest or Gradient Boosting.
• Develop a web application for real-time price prediction.

Acknowledgments 🙇

• Kaggle Dataset for providing the diamond dataset.
• Various Python libraries and tools used in the project.

License 📜

This project is licensed under the MIT License.