Skip to content

This project builds a predictive system for Interconnect, a telecommunications operator, to anticipate customer churn. Using historical data on contracts, services, and customer characteristics, the model will identify patterns to proactively manage customer retention strategies.

Notifications You must be signed in to change notification settings

Lani-Dom/Supervised-Learning-FinalProject

Repository files navigation

Supervised Learning (Final Project)

Churn Predictive Modeling

This project builds a predictive system for Interconnect, a telecommunications operator, to anticipate customer churn. Using historical data on contracts, services, and customer characteristics, the model will identify patterns to proactively manage customer retention strategies.

Highlights: data cleaning, exploratory data analysis, feature engineering, handling class imbalance, model selection (Logistic Regression, KNN Classifier, Random Forest, Gradient Boosting, XGBoost, LightGBM), hyperparameter tuning, performance metrics (AUC-ROC, Accuracy, Precision, Recall, F1-Score), and final model evaluation.

Installation

To install the project from GitHub, follow these steps:

  1. Clone the repository: git clone https://github.com/Lani-Dom/Supervised-Learning-FinalProject.git
  2. Navigate to the project directory: cd Supervised-Learning-FinalProject
  3. Install dependencies using the Requirements.txt file.
  4. Open the Jupyter Notebook to explore and execute the code.
  5. Utilize the datasets folder to access the data used in the project.

*This project was created by Lani Domínguez (Product Designer and Data Scientist) during TripleTen Academy's Data Science bootcamp.




Aprendizaje Supervisado (Proyecto Final)

Modelado Predictivo para Retención de Clientes

Este proyecto desarrolla un sistema predictivo para Interconnect, un operador de telecomunicaciones, con el objetivo de anticipar la cancelación de clientes. Utilizando datos históricos de contratos, servicios y características de los clientes, el modelo identificará patrones que ayuden a gestionar proactivamente la retención de clientes.

Características destacadas: limpieza de datos, análisis exploratorio de datos, ingeniería de características, manejo del desequilibrio de clases, selección de modelos (Regresión Logística, KNN, Bosques Aleatorios, Gradient Boosting, XGBoost, LightGBM), ajuste de hiperparámetros, métricas de rendimiento (AUC-ROC, Precisión, Recall, F1-Score), y evaluación final del modelo.

Instalación

Para instalar el proyecto desde GitHub, sigue estos pasos:

  1. Clona el repositorio: git clone https://github.com/Lani-Dom/SupervisedLearning-FinalProject.git
  2. Navega al directorio del proyecto: cd SupervisedLearning-FinalProject
  3. Instala las dependencias utilizando el archivo Requirements.txt.
  4. Abre el Jupyter Notebook para explorar y ejecutar el código.
  5. Utiliza la carpeta de datasets para acceder a los datos utilizados en el proyecto.

*Este proyecto fue creado por Lani Domínguez (Product Designer y Data Scientist) durante el bootcamp de Ciencia de Datos de TripleTen Academy.

About

This project builds a predictive system for Interconnect, a telecommunications operator, to anticipate customer churn. Using historical data on contracts, services, and customer characteristics, the model will identify patterns to proactively manage customer retention strategies.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published