import pandas as pd
import matplotlib.pyplot as plt
import numpy as np
import seaborn as sns

df = pd.read_csv('Datasets/heart_failure_clinical_records_dataset.csv')

df.head()

# Let's get a summary of our data set
df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 299 entries, 0 to 298
Data columns (total 13 columns):
 #   Column                    Non-Null Count  Dtype  
---  ------                    --------------  -----  
 0   age                       298 non-null    float64
 1   anaemia                   299 non-null    int64  
 2   creatinine_phosphokinase  292 non-null    float64
 3   diabetes                  294 non-null    float64
 4   ejection_fraction         294 non-null    float64
 5   high_blood_pressure       296 non-null    object 
 6   platelets                 288 non-null    object 
 7   serum_creatinine          299 non-null    float64
 8   serum_sodium              294 non-null    float64
 9   sex                       298 non-null    float64
 10  smoking                   298 non-null    float64
 11  time                      297 non-null    float64
 12  DEATH_EVENT               299 non-null    int64  
dtypes: float64(9), int64(2), object(2)
memory usage: 30.5+ KB

# Check for duplicate
df.duplicated().sum()

0

# Let's check how correlated each feature is to another
correlation = df.corr()
correlation

df.describe()

# Check number of columns and rows
df.shape #This shows that our dataset containes 299 rows and 13 columns

(299, 13)

# Let's checkout the column names
j = 0
for i in df.columns:
    j+=1
    print(j, i.upper())

1 AGE
2 ANAEMIA
3 CREATININE_PHOSPHOKINASE
4 DIABETES
5 EJECTION_FRACTION
6 HIGH_BLOOD_PRESSURE
7 PLATELETS
8 SERUM_CREATININE
9 SERUM_SODIUM
10 SEX
11 SMOKING
12 TIME
13 DEATH_EVENT

df['age'].T.hist(bins=50);

# Let's Visualize our correlations better
# sns.set(font_scale=1)
fig, ax = plt.subplots(figsize=(15,10))
ax = sns.heatmap(correlation, annot=True, fmt='.2f', cmap='YlGnBu', linewidths=.05);

pd.crosstab(df.age, df.DEATH_EVENT)

.dataframe tbody tr th:only-of-type {
    vertical-align: middle;
}

.dataframe tbody tr th {
    vertical-align: top;
}

.dataframe thead th {
    text-align: right;
}

df.isna()

.dataframe tbody tr th:only-of-type {
    vertical-align: middle;
}

.dataframe tbody tr th {
    vertical-align: top;
}

.dataframe thead th {
    text-align: right;
}

df.isna().sum()

age                          1
anaemia                      0
creatinine_phosphokinase     7
diabetes                     5
ejection_fraction            5
high_blood_pressure          3
platelets                   11
serum_creatinine             0
serum_sodium                 5
sex                          1
smoking                      1
time                         2
DEATH_EVENT                  0
dtype: int64

list(df.columns)

['age',
 'anaemia',
 'creatinine_phosphokinase',
 'diabetes',
 'ejection_fraction',
 'high_blood_pressure',
 'platelets',
 'serum_creatinine',
 'serum_sodium',
 'sex',
 'smoking',
 'time',
 'DEATH_EVENT']

# Split the data into features and labels
X = df.drop('DEATH_EVENT', axis=1)
y = df['DEATH_EVENT']

# We can use the SimpleImputer Class in scikit learn to do this but
from sklearn.impute import SimpleImputer
from sklearn.compose import ColumnTransformer


# Let's Define the columns to fill
categoricals = ['platelets', 'high_blood_pressure']
numerical = ['age',
             'anaemia',
             'creatinine_phosphokinase',
             'diabetes',
             'ejection_fraction',
             'serum_creatinine',
             'serum_sodium',
             'sex',
             'smoking',
             'time'] # Removed the two categorical feature (high_blood_pressure and platelets)

# Define the way SimpleImputer will fill the missing values
categorical_imputer = SimpleImputer(strategy='constant', fill_value='None')
numerical_imputer = SimpleImputer(strategy='mean')

# Create the Imputer 
transformer = ColumnTransformer([('categoricals', categorical_imputer, categoricals),
                                 ('numerical', numerical_imputer, numerical)])

new_X = transformer.fit_transform(X)

from sklearn.ensemble import RandomForestClassifier

# Instantiate our model classes
random_forest_model, ada_boost = RandomForestClassifier(), AdaBoostClassifier()

# Split our data into train and test splits
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(new_X, y, train_size=.8, random_state=21)

# Let's Checkout the shape of our data to understand how our data is being split
X_train.shape, X_test.shape, y_train.shape, y_test.shape

((239, 12), (60, 12), (239,), (60,))

random_forest_model.fit(X_train, y_train)

---------------------------------------------------------------------------

ValueError                                Traceback (most recent call last)

<ipython-input-306-7b70e76ee335> in <module>
----> 1 random_forest_model.fit(X_train, y_train)


/opt/anaconda3/lib/python3.8/site-packages/sklearn/ensemble/_forest.py in fit(self, X, y, sample_weight)
    302                 "sparse multilabel-indicator for y is not supported."
    303             )
--> 304         X, y = self._validate_data(X, y, multi_output=True,
    305                                    accept_sparse="csc", dtype=DTYPE)
    306         if sample_weight is not None:


/opt/anaconda3/lib/python3.8/site-packages/sklearn/base.py in _validate_data(self, X, y, reset, validate_separately, **check_params)
    431                 y = check_array(y, **check_y_params)
    432             else:
--> 433                 X, y = check_X_y(X, y, **check_params)
    434             out = X, y
    435 


/opt/anaconda3/lib/python3.8/site-packages/sklearn/utils/validation.py in inner_f(*args, **kwargs)
     61             extra_args = len(args) - len(all_args)
     62             if extra_args <= 0:
---> 63                 return f(*args, **kwargs)
     64 
     65             # extra_args > 0


/opt/anaconda3/lib/python3.8/site-packages/sklearn/utils/validation.py in check_X_y(X, y, accept_sparse, accept_large_sparse, dtype, order, copy, force_all_finite, ensure_2d, allow_nd, multi_output, ensure_min_samples, ensure_min_features, y_numeric, estimator)
    812         raise ValueError("y cannot be None")
    813 
--> 814     X = check_array(X, accept_sparse=accept_sparse,
    815                     accept_large_sparse=accept_large_sparse,
    816                     dtype=dtype, order=order, copy=copy,


/opt/anaconda3/lib/python3.8/site-packages/sklearn/utils/validation.py in inner_f(*args, **kwargs)
     61             extra_args = len(args) - len(all_args)
     62             if extra_args <= 0:
---> 63                 return f(*args, **kwargs)
     64 
     65             # extra_args > 0


/opt/anaconda3/lib/python3.8/site-packages/sklearn/utils/validation.py in check_array(array, accept_sparse, accept_large_sparse, dtype, order, copy, force_all_finite, ensure_2d, allow_nd, ensure_min_samples, ensure_min_features, estimator)
    614                     array = array.astype(dtype, casting="unsafe", copy=False)
    615                 else:
--> 616                     array = np.asarray(array, order=order, dtype=dtype)
    617             except ComplexWarning as complex_warning:
    618                 raise ValueError("Complex data not supportedn"


/opt/anaconda3/lib/python3.8/site-packages/numpy/core/_asarray.py in asarray(a, dtype, order, like)
    100         return _asarray_with_like(a, dtype=dtype, order=order, like=like)
    101 
--> 102     return array(a, dtype, copy=False, order=order)
    103 
    104 


ValueError: could not convert string to float: 'None'

# Let's checkout the unique values in our X,y set
y.unique(), new_X[::-1]

(array([1, 0]),
 array([['None', '0', 50.0, ..., 1.0, 1.0, 285.0],
        ['None', '0', 45.0, ..., 1.0, 1.0, 280.0],
        ['None', '0', 45.0, ..., 0.0, 0.0, 278.0],
        ...,
        ['162000', '0', 65.0, ..., 1.0, 1.0, 7.0],
        ['263358.03', '0', 55.0, ..., 1.0, 0.0, 6.0],
        ['265000', '1', 75.0, ..., 1.0, 0.0, 4.0]], dtype=object))

def fix_object_data(X):
    """
        This function will help us fix the string value errors in our dataset by turning string dtypes to numeric
        X: DataFrame or dict
    """
    for label, content in X.items():
        if pd.api.types.is_string_dtype(content):
            X[label] = pd.to_numeric(X[label], errors='coerce')
    return X

X.dtypes

age                         float64
anaemia                       int64
creatinine_phosphokinase    float64
diabetes                    float64
ejection_fraction           float64
high_blood_pressure          object
platelets                    object
serum_creatinine            float64
serum_sodium                float64
sex                         float64
smoking                     float64
time                        float64
dtype: object

fix_object_data(X).dtypes

age                         float64
anaemia                       int64
creatinine_phosphokinase    float64
diabetes                    float64
ejection_fraction           float64
high_blood_pressure         float64
platelets                   float64
serum_creatinine            float64
serum_sodium                float64
sex                         float64
smoking                     float64
time                        float64
dtype: object

X = fix_object_data(X)

def value_counts(data:pd.DataFrame, key='age'):
    columns = data.columns
    val_count = {}

    # Loop throught our data 
    for col in columns:
        val_count[col] = data[col].value_counts()
    return pd.DataFrame(val_count[key])

# Now let's consider filling this values one after the other since our features are not much
X['age'] = X['age'].fillna(X['age'].mean())
X['creatinine_phosphokinase'].fillna(X['creatinine_phosphokinase'].median(), inplace=True)
X['diabetes'].fillna(X['diabetes'].median(), inplace=True)
X['ejection_fraction'].fillna(X['ejection_fraction'].mean(), inplace=True)
X['high_blood_pressure'].fillna(X['high_blood_pressure'].mean(), inplace=True)
X['platelets'].fillna(X['platelets'].mean(), inplace=True)
X['serum_sodium'].fillna(X['serum_sodium'].median(), inplace=True)
X['sex'].fillna(X['sex'].median(), inplace=True)
X['smoking'].fillna(X['smoking'].median(), inplace=True)
X['time'].fillna(X['time'].median(), inplace=True)

X.isna().sum()

age                         0
anaemia                     0
creatinine_phosphokinase    0
diabetes                    0
ejection_fraction           0
high_blood_pressure         0
platelets                   0
serum_creatinine            0
serum_sodium                0
sex                         0
smoking                     0
time                        0
dtype: int64

from sklearn.model_selection import train_test_split
from sklearn.ensemble import RandomForestClassifier

# Split data into train and test sets
X_train, X_test, y_train, y_test = train_test_split(X,y, train_size=.8, random_state=21)

# Now let's train our train our model
model = RandomForestClassifier(n_estimators=120,max_depth=10)
model.fit(X_train, y_train)

RandomForestClassifier(max_depth=10, n_estimators=120)

model.predict(X_test)

array([1, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 1, 0, 0, 0,
       0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0,
       1, 0, 1, 0, 0, 0, 0, 1, 1, 1, 0, 1, 0, 0, 0, 0])

# Have a look at what our X_test looks like
X_test.head()

.dataframe tbody tr th:only-of-type {
    vertical-align: middle;
}

.dataframe tbody tr th {
    vertical-align: top;
}

.dataframe thead th {
    text-align: right;
}

from sklearn.metrics import confusion_matrix, accuracy_score, precision_score

y_pred = model.predict(X_test)
accuracy_score(y_test, y_pred)

0.9

import pickle as pkl

# Export with pickle
pkl.dump(model, open('your_first_model_pkl.pkl','wb')) #wb -> write binary

# Load with pickle
loaded_model_pkl = pkl.load(open('your_first_model_pkl.pkl', 'rb')) #rb -> read binary

# Make prediction
loaded_model_pkl.predict(X_test)

array([1, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 1, 0, 0, 0,
       0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0,
       1, 0, 1, 0, 0, 0, 0, 1, 1, 1, 0, 1, 0, 0, 0, 0])

import joblib as jbl

# Export with joblib
jbl.dump(model, 'your_first_model_pkl.joblib')

# Load with joblib
loaded_model_jbl = jbl.load('your_first_model_pkl.joblib')

# Make prediction
loaded_model_jbl.predict(X_test)

array([1, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 1, 0, 0, 0,
       0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0,
       1, 0, 1, 0, 0, 0, 0, 1, 1, 1, 0, 1, 0, 0, 0, 0])