Methodology - Processed Data

Outputs of the Project Methodology assignment, building on dallas_last_mile_dataset.csv from the Data Preparation stage (Processed Data dataset). dallas_methodology_dataset.csv adds the new accessibility_score column (equity-weighted composite: 0.5·z(resource_density) − 0.25·z(poverty_rate) − 0.25·z(zero_vehicle_hh_pct)) and drops the redundant *_count columns. regression_table*.csv/.arff and clustering_table*.csv/.arff are feature-selected working tables (_selected versions have the final chosen features only) used for the regression and clustering models respectively. The *_result.txt files are raw Weka output (feature selection, clustering, classifier evaluation) documenting exact settings and results. cluster_assignments.csv gives each station's K-Means and EM cluster label alongside its accessibility_score. pca_loadings.csv gives PCA component loadings from the unsupervised feature selection step. Software used: Weka 3.8.6 (command-line and GUI) and Python 3 (scikit-learn, pandas). Same 46-station Dallas DART walkshed study area as Data Preparation.