SpaceX Falcon 9 Launch Analysis

Published:

Role: Independent project · Period: Nov 2025

View on GitHub →

Overview

A complete data-science workflow on SpaceX Falcon 9 launch data — from raw records through cleaned features, predictive models, and interactive dashboards. The goal: identify the factors that drive launch success and booster-landing outcomes, and present the findings in a way a non-technical stakeholder could explore on their own.

What I did

  • Cleaned and transformed raw Falcon 9 launch records into an analysis-ready dataset.
  • Engineered features including mission type, payload mass ranges, and booster reuse counts to expose the operational variables most likely to drive outcomes.
  • Performed EDA across launch sites and conditions to surface patterns in success rates, landing types, and booster reuse.
  • Built and benchmarked machine learning models — Logistic Regression, Decision Tree, and SVM — to predict launch success, evaluating with accuracy, precision, recall, and confusion matrices.
  • Created interactive Plotly/Dash dashboards to visualize launch trends, landing success rates, and booster reuse performance — letting a viewer slice the data themselves rather than just consuming static charts.

Why it matters

The aerospace and engineering data science world increasingly relies on the same toolkit (cleaning → EDA → modeling → interactive visualization) as fintech and retail. This project is a concrete demonstration of that workflow on a public, real-world dataset, with a polished interactive deliverable rather than just a notebook.

Tech stack

Python · pandas · scikit-learn · IBM Watson Studio · SQL · Plotly · Dash · Logistic Regression · Decision Tree · SVM