Applied Data Science Capstone — SpaceX Falcon 9 Landing Prediction

IBM Data Science Professional Certificate — final capstone project.

Goal: predict whether the Falcon 9 first stage will land successfully, so launch cost can be estimated (a reused first stage costs ~$62M vs ~$165M for competitors).

Project stages

1. Data Collection — SpaceX REST API + Wikipedia web scraping (BeautifulSoup)
2. Data Wrangling — cleaning, one-hot encoding, building the binary landing label
3. EDA — SQL queries + visual exploration (matplotlib / seaborn)
4. Interactive Analytics — Folium maps + Plotly Dash dashboard
5. Predictive Analysis — Logistic Regression, SVM, Decision Tree, KNN with GridSearchCV

Key results

• Overall landing success rate ≈ 67% (60 successes / 30 failures over 90 launches)
• Success improves with flight experience and varies by orbit and launch site
• All four tuned classifiers reach ~83% accuracy on the test set

Contents

• dataset_part_2.csv — cleaned launch dataset with the landing Class label
• images/ — EDA, Folium and predictive-analysis charts used in the presentation
• notebooks/ — capstone notebooks
• SpaceX_Capstone_Presentation.pdf — final findings presentation