Deploy Machine Learning models using Flask into Heroku

import pandas as pdimport numpy as npimport picklefrom sklearn.linear_model import LogisticRegressionfrom sklearn.model_selection import train_test_split# Reading dataframe and dropping rows with na valuesdata = pd.read_csv("framingham.csv")data.dropna(inplace=True)  # Consists 3658 records# Computing Correlationcorr_matrix = data.corr().abs()high_corr_var = np.where(corr_matrix > 0.35)high_corr_var = [(corr_matrix.index[x], corr_matrix.columns[y]) for x, y in zip(*high_corr_var) if x != y and x < y]"""Variables to considerage: Age of a person (Input a number)smoker: Yes or NoCigs per day: (Input a number)diabaties: Yes or Nobmi: weight(Kg) and height(meters) calculateBP: input a number"""def bmi(weight, height):    return round(float(weight) / (float(height) * float(height)), 2)X_cols = ['male', 'age', 'currentSmoker', 'cigsPerDay', 'diabetes',          'sysBP', 'BMI']Y_col = ['TenYearCHD']X_vars = data[X_cols]Y_var = data[Y_col]# Renaming ColumnsX_vars.columns = ['Gender', 'Age', 'Smoker', 'Cigarettes_Per_Day',                  'Diabetic', 'BP', 'BMI']Y_var.columns = ['Chances_of_hear_disease']# Splitting dataX_train, X_test, y_train, y_test = train_test_split(X_vars, Y_var, test_size=0.25, random_state=0)# Initiate the Modellogreg = LogisticRegression()# fit the model with datalogreg.fit(X_train, y_train)pickle.dump(logreg, open('model.pkl', 'wb'))

# Set the secret key to some random bytes. # Keep this really secret!app.secret_key = b'_5#y2L"F4Q8z\n\xec]/'

import osimport picklefrom flask_wtf import FlaskFormfrom flask import Flask, render_templatefrom wtforms.validators import DataRequiredfrom wtforms import IntegerField, BooleanField, FloatField, SubmitField, RadioFieldclass Config(object):    SECRET_KEY = os.environ.get('SECRET_KEY') or 'you-will-never-guess'    APP_PATH = os.path.dirname(__file__)app = Flask(__name__)app.config.from_object(Config)class DataForm(FlaskForm):    gender = RadioField('Gender', validators=[DataRequired()], choices=['Male', 'Female'])    age = IntegerField('Age', validators=[DataRequired()])    diabetic = BooleanField('Are you Diabetic?')    smoker = BooleanField('Do you smoke? ')    cig_count = FloatField('How many cigarettes do you smoke per day ?', default=0)    weight = FloatField('Weight (Kg)', validators=[DataRequired()])    height = FloatField('Height (cm)', validators=[DataRequired()])    bp = IntegerField('Blood Pressure (mmHg)  Normal range : (90-120)/(60-80)', validators=[DataRequired()])    submit = SubmitField('Submit')@app.route('/')@app.route('/index', methods=['GET', 'POST'])def index():    form = DataForm()    if form.validate_on_submit():        li = ['age', 'gender', 'cig_count', 'diabetic', 'height', 'weight', 'smoker', 'bp']        data = {}        for ele in li:            data[ele] = eval('form.' + ele + '.data')        gender = 1 if data['gender'] == 'Male' else 0        smoker = 1 if data['smoker'] else 0        diabetic = 1 if data['diabetic'] else 0        filename = os.path.join(app.config['APP_PATH'], 'model.pkl')        with open(filename, 'rb') as f:            model = pickle.load(f)            value = model.predict([[gender, data['age'], smoker, data['cig_count'],                                    diabetic, data['bp'], bmi(data['weight'], data['height'])]])            data = ['done', value[0]]        return render_template('index.html', value=data, form=form, data=data)    return render_template('/index.html', form=form)def bmi(weight, height):    return round(float(weight) / (float(height / 100) * float(height / 100)), 2)if __name__ == '__main__':    app.run(debug=True)

web: gunicorn app:app

click==7.1.2Flask==1.1.2Flask-WTF==0.14.3gunicorn==20.0.4itsdangerous==1.1.0Jinja2==2.11.2joblib==0.16.0MarkupSafe==1.1.1numpy==1.19.1pandas==1.1.0python-dateutil==2.8.1python-dotenv==0.14.0pytz==2020.1scikit-learn==0.23.2scipy==1.5.2six==1.15.0sklearn==0.0threadpoolctl==2.1.0Werkzeug==1.0.1WTForms==2.3.3