The workflow begins by importing housing data using the Read CSV node. Outliers are detected and filtered using Flag Outlier and Row Filter nodes to ensure data quality. Categorical features are converted into numeric form with String Indexer and One Hot Encoder, followed by feature assembly through Vector Assembler.

The dataset is split into training and testing sets using the Split node. A Random Forest Regression model is then trained on the prepared data to predict housing prices. The Predict node applies the model to test data, and results are evaluated using the Regression Evaluator to measure performance.

The trained model is saved with ML Model Save for future deployment, and results are displayed using Drop Columns and Print N Rows. This end-to-end workflow enables scalable, accurate housing price prediction using distributed Spark ML processing.