For this project, my role as a data scientist was to analyze the commercial space industry. The focus is on predicting the successful reuse of the first stage.
By employing several different methodologies, I aim to predict the likelihood of first stage reuse. These predictions have implications for determining launch pricing and advancing reusable rocket technology
Methodology
- Data collection
- API integration and web scraping to collect relevant information on SpaceX, rocket launches, and first stage landings
- Data wrangling
- Calculations were made to identify the number of launches at each site, the number and occurrence of each orbit, the number and occurrence of mission outcome per orbit type, and created a landing outcome label from the outcome data
- EDA
- Performed exploratory data analysis using visualization and SQL
- Data Visualization
- Performed interactive visual analytics using Folium and Plotly Dash
- Performed predictive analysis using classification models
- Machine learning used to determine the first stage of Falcon 9 landing outcome. Split data into training data and test data to find the best Hyperparameter for SVM, Classification Trees, and Logistic Regression