Exploratory data analysis is the correct option because the team computed summary statistics, created histograms and box plots, and generated a correlation heatmap to understand distributions and relationships before cleaning or modeling.

These steps are classic exploratory data analysis tasks and help reveal patterns, outliers, and correlations that guide subsequent decisions about cleaning and feature design. EDA provides the domain view and questions that inform how to preprocess data and which features to create.

Data preprocessing is incorrect because it mainly refers to cleaning operations such as imputing missing values, encoding categorical variables, and scaling features rather than broad profiling and visualization.

Feature engineering is incorrect because it focuses on creating or transforming input variables to improve model performance and not on initial visual exploration and summary statistics.

Hyperparameter tuning is incorrect because it takes place after a model and features are in place and involves searching for optimal model settings like learning rate or tree depth.

Full AWS Practitioner Certification Question