Stellar Characteristics of the Solar Neighborhood: Machine Learning Insights for Space Navigation from Gaia Data Release 3

This study analyzes stars in the solar neighborhood — defined to be a heliocentric sphere with a 500 parsec radius — using stellar data from the Gaia Data Release 3 database. To allow for a feasible dataload with our computing power constraints, we restricted the stars by apparent magnitude, solely including stars within 7.5 <= G <= 13.0, where G is Gaia G-band (apparent) magnitude. We derived our restriction criteria and definition of the solar neighborhood from previous research by Zari et al 2018 and Prisinzano et al 2022. Using the columns for right ascension, declination, parallax, parallax-over-error, G magnitude, BP-RP color, proper motion right ascension, proper motion declination, and radial velocity, we construct interactive 3D spatial and kinematic maps, a luminosity function, and several machine-learning models: Isolation Forest for kinematic anomaly detection, a mass-estimation pipeline (analytic + Random Forest regression), and clustering in combined position–velocity space. Density-based clustering (HDBSCAN) initially returned 49 leaf clusters but with poor separation (silhouette –0.551), motivating a Gaussian Mixture fallback. Scanning k = 2–30 by BIC selected 20 Gaussian components as the preferred segmentation of the 500 pc neighborhood. Kinematic anomaly detection using an Isolation Forest (contamination 0.5%) flags a small, high-velocity subset for follow-up. Insights into local structure and motions may support future stellar navigation systems.