Car Price Prediction

This project is a car price prediction tool built with Python and Streamlit. The model leverages machine learning techniques to predict the price of a car based on various input features like model year, mileage, engine type, and more.

Project Owner

This project is a car price prediction tool built with Python and Streamlit. The model leverages machine learning techniques to predict the price of a car based on various input features like model year, mileage, engine type, and more.

Link to Streamlit App

Open in Streamlit

Features

• Predicts car prices based on user-provided details
• Interactive UI built with Streamlit for easy usability
• Data preprocessing, transformation, and model loading handled in the backend

Project Structure

project-folder/
│
├── main.py                  # Main Streamlit application
├── car_price_model.joblib   # Trained machine learning model
├── requirements.txt         # Python dependencies
└── README.md                # Project documentation

Installation

Clone the repository:

git clone https://github.com/yourusername/car-price-prediction.git
cd car-price-prediction

Install the required dependencies:

pip install -r requirements.txt

Run the Streamlit app:

streamlit run main.py

Open the app in your browser by going to http://localhost:8501.

Usage

1. Enter details like car model, year, mileage, transmission type, and other relevant information.
2. Click the "Predict Price" button to get an estimated price.

Model Training

The model was trained using various features, including:

• Model year
• Mileage
• Transmission type
• Fuel type
• Engine capacity and more...

The model was saved using joblib and can be reloaded for predictions.

Dependencies

This project uses the following libraries:

• Streamlit - For building the interactive web application
• Scikit-learn - For machine learning model training and pipeline management
• XGBoost - For the prediction model
• Pandas - For data manipulation and preprocessing

Contributing

Contributions are welcome! Please fork the repository and submit a pull request.

License

This project is licensed under the MIT License. See the LICENSE file for more details.

Notes

• Replace username in git clone with your GitHub username.

• If you don't have the requirements.txt file, you can create it by running the following command:

  pip freeze > requirements.txt

Code Explanation

Importing Libraries

import streamlit as st
import pandas as pd
import joblib

• streamlit: For creating an interactive UI.
• pandas: For data manipulation.
• joblib: For loading the saved model.

Setting up Streamlit UI

st.markdown("""
    <h1 style='text-align: right;'>مدل پیش‌بینی قیمت خودرو</h1>
    """, unsafe_allow_html=True)

• Displays the app title in Persian.

st.markdown("""
    <style>
    .css-1dq8tca {
        text-align: right;
        direction: rtl;
    }
    .stSelectbox, .stNumberInput, .stButton > button {
        text-align: right;
        direction: rtl;
        float: right;
    }
    </style>
    """, unsafe_allow_html=True)

• Sets up right-aligned UI elements for better Persian language support.

Loading Model and Data

train_data = pd.read_csv('train.csv')
model = joblib.load('car_price_model.joblib.gz')
preprocessor = joblib.load("car_price_model_preprocessor.joblib")

• train_data: Loads the dataset used for training.
• model: Loads the pre-trained model.
• preprocessor: Loads the preprocessor used to transform input data.

Encoding Categorical Features

column_dictionaries = {}
for col in train_data.select_dtypes(include='object').columns:
    unique_values = train_data[col].unique()
    value_to_number = {value: idx for idx, value in enumerate(unique_values)}
    column_dictionaries[col] = value_to_number

• Creates mappings for categorical values to convert text into numerical values.

Getting User Input

title = st.selectbox("نام خودرو", list(column_dictionaries['title'].keys()))
year = st.number_input("سال ساخت", min_value=1350, max_value=1403, value=1390)
mileage = st.number_input("کارکرد (به کیلومتر)", min_value=0.0, value=110000.0, step=1000.0)

• Collects user input for the car details, including title, year, and mileage.

Preparing Input Data

input_data = pd.DataFrame({
    "title": [column_dictionaries['title'][title]],
    "year": [year],
    "mileage": [mileage],
    "transmission": [column_dictionaries['transmission'][transmission]],
    "fuel": [column_dictionaries['fuel'][fuel]],
    "body_color": [column_dictionaries['body_color'][body_color]],
    "inside_color": [column_dictionaries['inside_color'][inside_color]],
    "body_status": [column_dictionaries['body_status'][body_status]],
    "body_type": [column_dictionaries['body_type'][body_type]],
    "volume": [volume],
    "engine": [column_dictionaries['engine'][engine]],
    "acceleration": [acceleration]
})

• Converts user inputs into a structured DataFrame, replacing categorical values with numerical representations.

Making Predictions

if st.button("پیش‌بینی قیمت"):
    predicted_price = model.predict(input_data)
    st.markdown(f"""
        <div style='text-align: right;'>
            <span style='font-weight: bold; font-size: 18px;'>قیمت پیش‌بینی شده: </span>
            <span style='color: purple; font-size: 24px; font-weight: bold;'>{predicted_price[0]:,.0f}</span>
            <span style='font-weight: bold; font-size: 18px;'>تومان</span>  
        </div>
    """, unsafe_allow_html=True)

• When the "Predict Price" button is clicked, the model predicts the price and displays it in Persian format.

---

Project link on my kaggle page

Project link