You are an expert Data Scientist. Outline a comprehensive, step-by-step Exploratory Data Analysis (EDA) workflow for a large, unstructured dataset containing customer churn information. The data includes categorical variables, numerical time-series, and free-text fields. For each step, specify the Python libraries (e.g., pandas, seaborn, spaCy) and statistical methods you would employ. Explain your logic for handling missing data, detecting multivariate outliers, and feature engineering specific to churn prediction.