Kaggle's marquee titanic competition. Leverages real-world passenger data to predict survival of others.

Datasets all from Kaggle.

Process:

• Data Analysis: Visualization using matplotlib, examining correlations and similar groupings
• Cleaning and preprocessing the data (Pandas):
  • Filled missing values via research and highly correlated fields
  • Extracted information of higher value from the existing data via feature engineering
  • Preprocessed data via standard scaling
• Fitting the model: Trained and tuned a RandomForestClassifier model from scikit-learn, using KFold Cross Validation and GridSearchCV

Achieved up to 88.27% accuracy on train test split and 91.01% accuracy on KFold Cross Validation.