Project Summary

Data science project analyzing SpaceX Falcon 9 launch outcomes using Python, SQL, Pandas, and NumPy. Built interactive visual analytics using Folium and Plotly Dash, and trained classification models with Scikit-learn to predict first-stage landing success. Includes data collection via APIs/web scraping, feature engineering, model evaluation, and visualization.

SpaceX Launch Analysis

End-to-end data science project analyzing Falcon 9 launches and predicting first-stage landing success.

Project Overview

This project explores SpaceX Falcon 9 launch data to understand what factors influence successful first-stage landings. I collected data from multiple sources (APIs and web scraping), cleaned and transformed it, performed exploratory data analysis, built interactive visualizations (maps and dashboards), and trained classification models to predict landing outcomes.

Key Highlights

Technical Stack

Data & Analysis

Python Pandas NumPy SQL

Visualization

Matplotlib Seaborn Plotly Folium

Dashboard

Plotly Dash

Machine Learning

Scikit-learn Logistic Regression SVM Decision Trees KNN

Leadership & Contributions

I completed this project end-to-end: collecting data, cleaning and engineering features, performing EDA, building interactive maps and dashboards, and training multiple machine learning models to compare performance.

Results

The analysis showed that variables such as payload mass, orbit type, and launch site are associated with landing outcomes. I compared multiple classification models and selected a best-performing approach based on test accuracy and overall classification metrics.

Deliverables include cleaned datasets, visual analytics, an interactive dashboard, and a reusable predictive modeling pipeline.

Key Insights

Visual Highlights

SpaceX visual analytics including launch site map markers, successful launches by site chart, and payload versus outcome plot
Visual analytics: launch site map outcomes, successful launches by site, and payload vs outcome exploration.
Confusion matrix evaluating classifier predictions for landing versus did not land
Confusion matrix used to evaluate classification model performance for predicting landing success.

Visual artifacts include interactive mapping concepts (Folium), dashboard-style charts, and model evaluation outputs (confusion matrix) used to validate classifier performance.

Links