Car Price Range Estimation App

This project is designed to estimate a car price range suitable for customers based on their financial information. By inputting details such as age, salary, and net worth, the app calculates and advises a car price range tailored to the customer.

Project Overview

The goal of this project is to provide a seamless interface for users to input their details and receive an estimated car price range based on their financial data. The app uses simple calculations and a clean user interface to deliver actionable insights.

Key Steps

1. Data Preparation

• Loaded the car purchasing data from a CSV file (car_purchasing.csv).

  import pandas as pd
  df = pd.read_csv("car_purchasing.csv", encoding="latin-1")

• Displayed basic dataset information including shape, column names, and data types using info() and shape.
• Inspected the first few rows using head().
• Checked for and handled missing values (if any) using isnull().sum() and potentially dropna().
• Checked for and removed duplicate entries (if any) using drop_duplicates().

2. Exploratory Data Analysis (EDA)

• Calculated descriptive statistics for numerical columns using describe().

• Analyzed the distribution of key features, including age, annual salary, credit card debt, and net worth, using histograms and scatter plots.

• Visualized relationships between features and the target variable (car purchase amount) using scatter plots.

• Calculated correlations between numerical features and the target variable using corr().

• Explored potential categorical features like gender using groupby() and visualizations.

3. Data Visualization

Scatter plots: Visualized the relationships between numerical features (age, salary, net worth, credit and debt) and car purchase amount.

Histograms: Showed the distribution of age in the dataset.

Bar plots: Compared mean net worth and Mean annual salary by gender, and visualized mean/maximum car purchase amount by age range.

3. Feature Engineering

• Created new features to potentially improve model performance:
  • Converted gender to a categorical variable (n_gender) using apply().

  df['n_gender'] = df['gender'].apply(lambda x: 1 if x == 'Male' else 0)

  • Grouped age into ranges (age_range) using a custom function and apply().

  def age_group(age):
      if age < 30:
          return 'Young'
      elif 30 <= age < 50:
          return 'Middle-aged'
      else:
          return 'Senior'
  df['age_range'] = df['age'].apply(age_group)

• Analyzed the impact of the new features on the target variable.

4. Model Building and Evaluation

• Split the dataset into training and testing sets using train_test_split from sklearn.model_selection.

  from sklearn.model_selection import train_test_split
  X = df[['age', 'annual_salary', 'credit_card_debt', 'net_worth']]
  y = df['car_purchase_amount']
  X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

• Scaled numerical features using StandardScaler from sklearn.preprocessing.

  from sklearn.preprocessing import StandardScaler
  scaler = StandardScaler()
  X_train = scaler.fit_transform(X_train)
  X_test = scaler.transform(X_test)

• Trained three regression models:
  • Linear Regression

  from sklearn.linear_model import LinearRegression
  lr = LinearRegression()
  lr.fit(X_train, y_train)

  • Support Vector Regression

  from sklearn.svm import SVR
  svr = SVR()
  svr.fit(X_train, y_train)

  • Random Forest Regression

  from sklearn.ensemble import RandomForestRegressor
  rf = RandomForestRegressor()
  rf.fit(X_train, y_train)

• Performed hyperparameter tuning for SVR and Random Forest Regression using GridSearchCV to optimize model performance.
• Evaluated model performance using metrics like Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE) using functions from sklearn.metrics.

5. Model Selection and Saving

• Selected the best-performing model based on evaluation results. (In this case, it was likely Linear Regression).
• Saved the chosen model using joblib.dump for future use.

  import joblib
  joblib.dump(lr, "model.pkl")
  joblib.dump(scaler, "scaler.pkl")

Streamlit Application

• A Streamlit app was developed to allow users to input customer details and get a recommended car price range.

  age =st.number_input('Age',min_value=18,max_value=90,value=40, step=1)
  annual_salary = st.number_input('Annual Salary', min_value=500, max_value=9999999999, value=12000, step=5000)
  net_worth = st.number_input('Net Worth', min_value=0, max_value=999999999, step=2000, value=100000)
  
  
  X = [age,annual_salary,net_worth]
  calculate =st.button('Calculate')
  st.divider()
  
  X_scaled = scaler.transform([X])
  
  if calculate:
    st.balloons()
    X_2 = np.array(X)
    X_array = scaler.transform([X_2])
    
    prediction = model.predict(X_array)
    
    # Check if prediction is less than 0
    if prediction[0] < 0:
        st.write("Prediction is: Negative")
    else:
        st.write(f"Prediction is: {prediction[0]:}")
    
    st.write("Advice: cars in the similar values")
       
  else:
    st.write("Enter values")

Results

• The application provides quick and accurate car price range estimates based on user inputs.
• Screenshot of the app interface:

Future Work

• Improve the model by incorporating more complex features.
• Compare performance with advanced models like Gradient Boosting or Neural Networks.
• Deploy the app on a cloud platform for wider accessibility.

Contributing

Contributions are welcome! Please follow these steps:

1. Fork the repository.
2. Create a new branch for your feature or bugfix:

   git checkout -b feature-name

3. Commit your changes:

   git commit -m "Description of your changes"

4. Push to your branch:

   git push origin feature-name

5. Create a pull request.

License

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