A Comprehensive Data-Driven Study of SpaceX Falcon 9 Launch Outcomes

This project presents a complete data science workflow analyzing real SpaceX Falcon 9 launch data. It covers everything from data collection and wrangling to exploratory data analysis, interactive visualization, and predictive modeling. Goal: Identify key factors influencing rocket landing success and build predictive models to estimate launch outcomes. Project Pipeline Phase Description 1. Data Collection Fetched launch data using the SpaceX REST API (v4) and Wikipedia scraping (BeautifulSoup). 2. Data Wrangling Cleaned, merged, and filtered datasets. Created the Class label (1 = success, 0 = failure). 3. Exploratory Data Analysis (EDA) Visualized success patterns across payload mass, orbit type, and launch site. 4. Interactive Visualization Built Folium maps for spatial exploration and a Plotly Dash dashboard for user-driven analysis. 5. Predictive Analytics Trained and evaluated ML models (Logistic Regression, SVM, Decision Tree, KNN) using GridSearchCV. Key Insights Payload mass has a non-linear relationship with landing success. Launch Site KSC LC-39A has the highest mission success rate. Decision Tree classifier achieved 90% accuracy with zero false negatives. SpaceX’s mission reliability shows a steady improvement across years. Future Enhancements Integrate recent launches beyond 2020 from the SpaceX API. Implement ensemble ML models (RandomForest, XGBoost). Deploy the Dash app publicly using Render or Heroku. Add a real-time API update feature to keep dashboard live.