# App of GenAI/Deep Learning (WUSTL, Spring 2026) — Data 来源：https://www.kaggle.com/competitions/app-of-gen-ai-deep-learning-wustl-spring-2026/data 采集时间：2026-10-19T08:01:11Z # Exoplanet Settlement Viability Challenge ## Competition Overview Welcome to the Exoplanet Settlement Viability Challenge! In this competition, you'll analyze data from newly discovered exoplanets evaluated by an interstellar exploration agency. Your task is to predict the **settlement viability index** for each planet based on a rich set of scientific measurements, environmental observations, and field reports. ## Evaluation Metric Submissions are evaluated using **Root Mean Squared Error (RMSE)** between the predicted settlement viability index and the actual values. $$\text{RMSE} = \sqrt{\frac{1}{n}\sum_{i=1}^{n}(y_i - \hat{y}_i)^2}$$ Where: - $n$ is the number of planets in the test set - $y_i$ is the actual settlement viability index - $\hat{y}_i$ is your predicted settlement viability index Lower RMSE values indicate better predictions. ## Dataset Description The dataset contains comprehensive information about exoplanets discovered during deep-space exploration missions. Each planet has been surveyed by robotic probes and landing teams, generating both structured scientific measurements and unstructured field observations. ### Files - **train.csv** - Training set with settlement viability index - **test.csv** - Test set without settlement viability index (your prediction target) - **sample_submission.csv** - Sample submission file in the correct format - **images/** - Directory containing exterior planetary surface images (named by planet ID, e.g., `1.jpg`, `2.jpg`, etc.) ### Images Each planet entry includes a high-resolution image of the planetary surface captured during initial survey missions. These images are named using the planet's ID (e.g., planet with `id=42` has image `42.jpg`). The images show wide-angle cinematic views of the surface environment and may provide additional visual context about atmospheric conditions, terrain features, and overall habitability. ## Data Dictionary | Column Name | Type | Description | | ----------------------------- | ------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **id** | integer | Unique identifier for each planet | | **planet_name** | string | Designated name of the exoplanet | | **star_name** | string | Name of the host star system | | **star_type** | string | Classification of the host star (e.g., red dwarf, G-type main sequence, K-type, binary) | | **planet_type** | string | Primary classification of the planet (e.g., terrestrial, ocean world, ice world, desert world, volcanic world, super-Earth, gas giant moon, tidally locked, rogue planet) | | **orbital_distance_au** | float | Distance from host star in astronomical units (AU) | | **orbital_period_days** | float | Time to complete one orbit around the host star in Earth days | | **day_length_hours** | float | Duration of one planetary rotation in Earth hours | | **axial_tilt_deg** | float | Axial tilt of the planet in degrees | | **mean_temp_c** | float | Mean surface temperature in degrees Celsius | | **temp_range_c** | float | Temperature variation between extremes in degrees Celsius | | **gravity_g** | float | Surface gravity relative to Earth (1.0 = Earth gravity) | | **escape_velocity_kms** | float | Minimum velocity needed to escape the planet's gravity in km/s | | **atmospheric_pressure_atm** | float | Surface atmospheric pressure relative to Earth (1.0 = Earth sea level) | | **atmosphere_primary_gases** | string | Composition of the atmosphere with percentages | | **oxygen_percent** | float | Percentage of oxygen in the atmosphere | | **co2_percent** | float | Percentage of carbon dioxide in the atmosphere | | **methane_percent** | float | Percentage of methane in the atmosphere | | **water_presence** | string | Classification of water availability (none/trace/subsurface/seas/global ocean) | | **surface_composition** | string | Primary surface material composition (rocky/metallic/sandy/ice/crystalline/volcanic) | | **dominant_biome** | string | Primary environmental biome classification | | **environment_type** | string | Overall environmental classification for settlement assessment | | **albedo** | float | Reflectivity of the planet's surface (0.0 = absorbs all light, 1.0 = reflects all light) | | **cloud_coverage_percent** | float | Percentage of surface covered by clouds | | **storm_frequency** | string | Frequency of severe weather events (rare/seasonal/frequent/constant) | | **seismic_activity** | string | Level of tectonic/seismic activity (low/moderate/high/extreme) | | **magnetic_field** | string | Strength of planetary magnetic field (weak/moderate/strong) | | **radiation_level** | string | Surface radiation exposure level (low/moderate/high/extreme) | | **moons_count** | integer | Number of natural satellites orbiting the planet | | **rings** | string | Presence of planetary ring system (yes/no) | | **native_biology** | string | Classification of indigenous life forms (none/microbial/complex/unknown) | | **invasive_risk** | integer | Risk assessment for biological contamination (0-10 scale, higher = more dangerous) | | **rare_mineral_index** | integer | Abundance of valuable mineral resources (0-100 scale, higher = more abundant) | | **energy_harvest_potential** | string | Potential for renewable energy generation (low/moderate/high/extreme) | | **terraforming_difficulty** | integer | Estimated difficulty of terraforming operations (0-10 scale, higher = more difficult) | | **settlement_cost_index** | integer | Estimated relative cost of establishing a settlement (0-100 scale, higher = more expensive) | | **strategic_value_rating** | integer | Strategic importance for expansion goals (0-100 scale, higher = more valuable) | | **primary_hazard** | string | Most significant danger to settlement efforts (toxic_air/radiation/predators/seismic/storms/unknown_anomaly) | | **viability_class** | string | Preliminary classification (Not Viable / Research Outpost Only / Partial Terraforming Required / Immediately Settleable) | | **target_habitability_score** | integer | Internal habitability assessment score (0-100 scale) | | **exploration_log** | string | First-landing field notes from survey team (200-400 words) | | **environmental_report** | string | Technical environmental assessment report (120-250 words) | | **incident_report** | string | Documentation of notable events during survey (60-150 words) | | **image_prompt** | string | Description of the planetary surface scene captured in the associated image | | **settlement_index** | float | **[TARGET]** Overall viability score for establishing a successful settlement (0-100 scale, higher = more viable) | ## Target Variable **settlement_index**: A comprehensive metric (0-100) that represents the overall suitability of a planet for establishing a successful, self-sustaining human settlement. This score integrates multiple factors including environmental conditions, resource availability, safety considerations, and economic feasibility. Higher scores indicate planets that are more suitable for settlement efforts. ## Notes - The dataset includes both structured numerical/categorical features and unstructured text fields - Missing values may exist in some columns and should be handled appropriately - The target variable is continuous and ranges from 0 to 100 - Text fields (exploration_log, environmental_report, incident_report) may contain valuable information for feature engineering - Planetary images are provided as supplementary data and may be used for additional feature extraction if desired - All measurements use Earth-standard units for reference Good luck, and may your models identify the most promising worlds for humanity's future!