Titanic-Data-Analysis-and-Prediction

Titanic Data : Exploratory Data Analysis & Survivor's Prediction

EDA Analysis:

• Check the Data + Correlation Maps + Check NaN values
• Question: Calculate the number of surviving/non-surviving passengers.
• Question: Calculate the proportion of surviving 1st class passengers with regards to the total number of 1st class passengers.
• Question: Create a bar plot with separate bars for male/female passengers and 1st/2nd/3rd class passengers.
• Question: Create Bar plot for survived and drowned passengers for different classes.
• Question: Classify Man/Woman/Child surviving and drowned.
• Question: Create a histogram showing the age distribution of passengers. Compare surviving/non-surviving passengers.
• Question: Calculate the average age for survived and drowned passengers separately.
• Question: Create a table counting the number of surviving/drowned passengers separately for 1st/2nd/3rd class and male/female.

Machine Learning

• Building model for the Titanic data and predicting the unseen data Logistic Regression model: First I choose the features and did a train_test_split. Applied feature engineering to different features. Calculated the accuracy scores. Created pickle files for future use. Finally, predicted the test/unseen data.
• Build a Baseline Model: To better understand the data baseline model is created. DummyClassifier: A classifier model that make predictions without trying to find patterns in the data. Calculate the accuracy, recall, precision and f1 score with different strategy.
• RandomForest: This is done to check the important features in our dataset.