emmanuelalo52/Telco-churn-prediction

Telco Customer Churn Prediction

This repository contains code for predicting customer churn in a telecommunications company using the CatBoost classifier. The project includes data preprocessing, model training, evaluation, and deployment using Streamlit and FastAPI.

Table of Contents

1. Installation
2. Data Preprocessing
3. Model Training
4. Evaluation
5. Deployment
6. Usage

Installation

To get started with this project, clone the repository and install the required dependencies:

git clone https://github.com/yourusername/telco-churn-prediction.git
cd telco-churn-prediction
pip install -r requirements.txt

Data Preprocessing

The data preprocessing steps include handling missing values, encoding categorical variables, and splitting the dataset into training and testing sets.

Steps:

1. Load Data:

   telecom_df = pd.read_csv(data_path)

2. Handle Missing Values:

   telecom_df["TotalCharges"] = pd.to_numeric(telecom_df["TotalCharges"], errors='coerce')
   telecom_df["TotalCharges"].fillna(telecom_df["tenure"] * telecom_df["MonthlyCharges"], inplace=True)

3. Encode Categorical Variables:

   telecom_df['SeniorCitizen'] = telecom_df['SeniorCitizen'].astype(object)
   columns_to_replace = ['OnlineSecurity', 'OnlineBackup', 'DeviceProtection', 'TechSupport', 'StreamingTV', 'StreamingMovies']
   for column in columns_to_replace:
       telecom_df[column] = telecom_df[column].replace('No internet service', 'No')
   telecom_df['Churn'] = telecom_df['Churn'].replace({'No': 0, 'Yes': 1})

4. Split Data:

   strasplit = StratifiedShuffleSplit(n_splits=1, test_size=0.2, random_state=42)
   train_index, test_index = next(strasplit.split(telecom_df, telecom_df['Churn']))
   strat_train = telecom_df.loc[train_index]
   strat_test = telecom_df.loc[test_index]
   x_train = strat_train.drop("Churn", axis=1)
   y_train = strat_train["Churn"].copy()
   x_test = strat_test.drop("Churn", axis=1)
   y_test = strat_test["Churn"].copy()

Model Training

The model training involves initializing the CatBoost classifier and fitting it to the training data.

Steps:

1. Identify Categorical Features:

   categorical_columns = telecom_df.select_dtypes(include=['object']).columns.to_list()

2. Initialize and Train the Model:

   catboost_model = CatBoostClassifier(verbose=False, random_state=0, scale_pos_weight=3)
   catboost_model.fit(x_train, y_train, cat_features=categorical_columns, eval_set=(x_test, y_test))

Evaluation

Evaluate the model using various metrics such as accuracy, recall, precision, and ROC-AUC.

Steps:

1. Predict and Calculate Metrics:

   y_pred = catboost_model.predict(x_test)
   accuracy, recall, roc_auc, precision = [round(metric(y_test, y_pred), 4) for metric in [accuracy_score, recall_score, roc_auc_score, precision_score]]

2. Store Results:

   model_name = ['Catboost_model']
   result = pd.DataFrame({'Accuracy': accuracy, 'Recall': recall, 'Roc_Auc': roc_auc, 'Precision': precision}, index=model_name)
   print(result)

3. Save the Model:

   model_path = os.path.join(model_dir, "catboost_model.cbm")
   catboost_model.save_model(model_path)

Deployment

The deployment includes creating a Streamlit app for user interaction and a FastAPI service for model inference.

Streamlit App

1. Load Model and Data:

   model = load_model()
   data = load_data()

2. User Interface for Churn Prediction:

   st.title("Telco Customer Churn Project")

3. SHAP Values for Model Explanation:

   plot_shap_values(model, explainer, shap_values_cat_train, shap_values_cat_test, customer_id, X_test, X_train)

FastAPI Service

1. Define API Endpoint:

   app = FastAPI(title="Churn prediction", version="1.0")

2. Churn Prediction Endpoint:

   @app.post('/predict/')
   def predict_churn(data: dict):
       churn_probability = get_churn_probability(data, model)
       return {'Churn probability': churn_probability}

3. Run the Service:

   uvicorn api:app --host 127.0.0.1 --port 5000

Usage

To use the churn prediction model, run the Streamlit app for an interactive interface or send POST requests to the FastAPI service with customer data to get churn probability predictions.

Streamlit App

Run the Streamlit app:

streamlit run app.py

FastAPI Service

Start the FastAPI service:

uvicorn api:app --host 127.0.0.1 --port 5000

Send a POST request to the /predict/ endpoint with customer data to get the churn probability.

{
  "customerID": "6464-UIAEA",
  "gender": "Female",
  "SeniorCitizen": 0,
  "Partner": "Yes",
  "Dependents": "No",
  "tenure": 12,
  "PhoneService": "Yes",
  "MultipleLines": "No",
  "InternetService": "DSL",
  "OnlineSecurity": "No",
  "OnlineBackup": "Yes",
  "DeviceProtection": "No",
  "TechSupport": "Yes",
  "StreamingTV": "No",
  "StreamingMovies": "Yes",
  "Contract": "Month-to-month",
  "PaperlessBilling": "Yes",
  "PaymentMethod": "Electronic check",
  "MonthlyCharges": 29.85,
  "TotalCharges": 29.85
}

This README provides a comprehensive guide to understanding, installing, and using the Telco Customer Churn Prediction project. Follow the steps to preprocess data, train the model, evaluate its performance, and deploy it using Streamlit and FastAPI.