Implementación de una estrategia de marketing basada en modelo supervisado para definir clientes propensos a retirar efectivo con tarjeta de crédito

Abstract

El presente proyecto describe una propuesta de implementación de una estrategia de mercadeo utilizando el poder de los datos para impulsar el servicio financiero de retiro de efectivo con tarjeta de crédito. Se utiliza un modelo de predicción clasificatorio para identificar a los clientes propensos a retirar efectivo por primera vez y a impulsar más retiros de los clientes ya existentes. Se emplea la metodología CRISP-DM de la ciencia de datos para detectar el problema, la solución y el modelo con mejor rendimiento para identificar a los segmentos de clientes, permitiendo diseñar una estrategia de marketing basada en datos de alto valor.

This project presents a proposal for implementing a marketing strategy that leverages the power of data to promote a financial service. The service targeted by the proposed segmentation strategy is credit card cash advances. Financial institutions manage large volumes of customer data to deliver their services, making customer segmentation a challenging task. Therefore, a supervised classification model was developed to identify customers who are likely to make their first credit card cash advance, as well as existing customers who are likely to increase their use of this service.

The project follows the Cross-Industry Standard Process for Data Mining (CRISP-DM) methodology to identify the business problem, define the proposed solution, and select the predictive model with the best performance for identifying customer segments. Based on the information collected, a data-driven marketing strategy was designed.

The findings of this project demonstrate that data mining provides significant value for the development of effective marketing strategies, enabling more accurate customer segmentation and supporting informed decision-making for financial service promotion.