This case study explores opportunities to increase the likelihood that customers recommend the company’s customer service. It is a continuation of the project [Predicting Customer Detractors (Part 1): Analyzing Contextual Factors via Logistic Regression](https://rpubs.com/galdovaldonavas/1335090), which identified the service contexts with the lowest likelihood of recommendation (i.e., contact methods, contact reasons, and countries). Building on those findings, this follow-up project focuses on **understanding the root causes of dissatisfaction** and quantifying the potential benefits of addressing them. Specifically, we conduct a thematic analysis of open-ended feedback from Net Promoter Score (NPS) surveys to identify actionable issues. We then assess the expected impact of solving these problems on customers’ likelihood to recommend. The analysis includes: - Data simulation and cleaning. - Exploratory analysis with descriptive statistics and visualizations. - Logistic regression modeling. - Simulation-based predictions using bootstrapping. Although based on a real-world project, all data, variables, and insights presented here have been simulated to ensure confidentiality.

This case study aims to identify key factors that influence customer's likelihood to recommend the company after interacting with customer service. Methodology: The project utilizes a comprehensive analytical approach, including: - Data Simulation & Cleaning: Creating and preparing the dataset for analysis. - Exploratory Data Analysis: Using data visualization (e.g., heatmaps) and descriptive statistics to uncover patterns across multiple and interactive factors. - Statistical Modeling: Evaluating different regression models (linear, ordinal, binomial) to predict customer's likelihood to recommend the company. - Simulation Based Recommendations: Predictions to evaluate the impact of different actions. - Reusable Functions: The creation of functions to automate procedures. Tools & Libraries: R with a focus on libraries such as car, VGAM, ordinal, psych, vcd, coefplot, ggplot2, tidyr, dplyr, openxlsx, and readxl.