Overview
SpaceX advertises Falcon 9 rocket launches at $62 million while competitors charge upwards of $165 million. The cost savings are largely due to reusable first stages. Predicting landing success is crucial for determining launch costs and competitiveness. This project analyzes historical launch data to predict whether SpaceX will attempt to land the first stage.
What I Did
•Data Collection & Cleaning
•Exploratory Data Analysis
•Feature Engineering
•Model Training & Evaluation
•Results Visualization
Challenges
- –Collecting and cleaning SpaceX launch data from multiple sources
- –Handling imbalanced dataset (more successful than failed landings)
- –Selecting optimal features for prediction accuracy
Solutions
- ✓Used web scraping and API calls to gather comprehensive data
- ✓Applied SMOTE and class weighting to handle imbalanced classes
- ✓Conducted thorough feature importance analysis and selection
Technologies
PythonPandasNumPyScikit-learnMatplotlibSeabornJupyter Notebook
Results
85%+
Model Accuracy
Successfully predicted landing outcomes
15+
Features Analyzed
Flight parameters and conditions
1000+
Data Points
Historical launch records processed