notha99y

plt.figure(figsize=(14,12))
plt.title('Pearson Correlation of Features', size = 15)
colormap = sns.diverging_palette(10, 220, as_cmap = True)
sns.heatmap(corr_df.corr(),
            cmap = colormap,
            square = True,
            annot = True,
            linewidths=0.1,vmax=1.0, linecolor='white',
            annot_kws={'fontsize':12 })
plt.show()

notha99y

from scipy.cluster.hierarchy import linkage
from scipy.cluster.hierarchy import dendrogram
sample_train,sample_val, gt_train, gt_val = train_test_split(train_df,
                                                             train_df['Survived'],
                                                             test_size=0.05,
                                                             random_state=99)

sample_val_processed = simple_preprocessing(sample_val, train = False)
sample_val_processed = scaler.fit_transform(sample_val_processed)
mergings = linkage(sample_val_processed, method='complete')

notha99y

def simple_preprocessing(dataframe, train=True):
    le = LabelEncoder()
    X = dataframe.drop(['PassengerId', 'Cabin', 'Name', 'Ticket'], axis=1)
    X['Age'] = X['Age'].fillna(value=X['Age'].mode()[0])
    X['Embarked'] = le.fit_transform(X['Embarked'].fillna(value=X['Embarked'].mode()[0]))
    X['Sex'] = np.where(X['Sex'] == 'male', 1, 0)
    
    if train:
        X = X.drop(['Survived'], axis=1)
        y = np.where(dataframe['Survived'] == 1, 'Alive', 'Dead')

notha99y

from xgboost import XGBClassifier

xgb_clf = XGBClassifier(max_depth=12, learning_rate=1e-4,n_estimators=500)
xgb_clf.fit(X_train, np.argmax(np.array(y_train), axis = 1))
xgb_y_pred = xgb_clf.predict(X_val)

pd.Series(xgb_clf.feature_importances_, index = X_train.columns).nlargest(12).plot(kind = 'barh',
                                                                               figsize = (10, 10),
                                                                              title = 'Feature importance from XGBoost').invert_yaxis();

notha99y

from sklearn.ensemble import RandomForestClassifier
rf_clf = RandomForestClassifier(n_estimators = 500, max_depth=12)
rf_clf.fit(X_train, y_train)
rf_y_pred = rf_clf.predict(X_val)

pd.Series(rf_clf.feature_importances_, index = X_train.columns).nlargest(12).plot(kind = 'barh',
                                                                               figsize = (10, 10),
                                                                              title = 'Feature importance from RandomForest').invert_yaxis();

notha99y

# multivariate analysis with Embarked variable and Pclass variable
quantitative_summarized(dataframe= train_df, y = 'Age', x = 'Embarked', hue = 'Pclass', palette=c_palette3, verbose=False, swarm=False)

notha99y

# bivariate analysis with target variable
quantitative_summarized(dataframe= train_df, y = 'Age', x = 'Survived', palette=c_palette, verbose=False, swarm=True)

notha99y

# univariate analysis
quantitative_summarized(dataframe= train_df, y = 'Age', palette=c_palette, verbose=False, swarm=True)

notha99y

def quantitative_summarized(dataframe, x=None, y=None, hue=None, palette='Set1', ax=None, verbose=True, swarm=False):
    '''
    Helper function that gives a quick summary of quantattive data

    Arguments
    =========
    dataframe: pandas dataframe
    x: str. horizontal axis to plot the labels of categorical data (usually the target variable)
    y: str. vertical axis to plot the quantitative data
    hue: str. if you want to compare it another categorical variable (usually the target variable if x is another variable)

notha99y

# Feature Variable: Gender
categorical_summarized(train_df, y = 'Sex', hue='Survived', palette=c_palette)