Predictive Analysis for SpaceX Falcon 9 First Stage Landing Success
In this capstone project, I undertook a comprehensive analysis as part of the IBM Data Science Specialization, focusing on predicting the successful landing of the Falcon 9 first stage in SpaceX rocket launches. The overarching goal was to provide insights into the cost-effectiveness of SpaceX launches, given the substantial savings achieved through stage reuse. Priced at $62 million, SpaceX launches present a formidable cost advantage compared to competitors, making accurate landing predictions crucial for companies in the competitive rocket launch market.
The project commenced by leveraging the SpaceX API to retrieve relevant data, followed by meticulous cleaning to ensure data accuracy and completeness. A series of exploratory data analysis (EDA) tasks were performed to gain insights into the dataset, and feature engineering techniques were applied to enhance predictive capabilities. The next steps involved standardizing the data, splitting it into training and test sets, and identifying the best hyperparameters for Support Vector Machines (SVM), Classification Trees, and Logistic Regression.
The ultimate objective was to determine the optimal predictive model by assessing performance on the test data. This predictive analysis is invaluable for companies bidding against SpaceX, providing them with a strategic advantage by anticipating the success of Falcon 9 first stage landings and subsequently estimating the cost-effectiveness of competing in the rocket launch market. The capstone project showcases a holistic approach to data science, encompassing data acquisition, cleaning, exploratory analysis, and predictive modeling, with real-world applications in the evolving space industry.
The data used for this project can be found here.
Skills & Tools
Data Science
, Predictive Models
, Machine Learning
, Exploratory Data Analysis
, Data Visualization
, Python
, Decision Trees
, Support Vector Machines
, Logistic Regression