app.py

import streamlit as st
import numpy as np
import pandas as pd
import pickle
from sklearn.preprocessing import  LabelEncoder
file_path = '/content/drive/MyDrive/CODING DOJO DS BOOTCAMP/PROJECTS/PROJECT 2/Stroke Prediction Dataset'

df = pd.read_csv('stroke-data.csv')

#load label encoder
le=LabelEncoder()

# Load our model
pickle_in = open('best_vc_ensemble.pkl', 'rb')
model = pickle.load(pickle_in)


def predict_stroke(gender, age, hypertension, heart_disease, ever_married,
       work_type, Residence_type, avg_glucose_level, bmi,
       smoking_status):
  le=LabelEncoder()

  le.fit(df['gender'].value_counts().index)
  gender = le.transform([f'{gender}'])[0]

  le.fit(df['ever_married'].value_counts().index)
  ever_married = le.transform([f'{ever_married}'])[0]

  le.fit(df['work_type'].value_counts().index)
  work_type = le.transform([f'{work_type}'])[0]

  le.fit(df['Residence_type'].value_counts().index)
  Residence_type = le.transform([f'{Residence_type}'])[0]

  le.fit(df['smoking_status'].value_counts().index)
  smoking_status = le.transform([f'{smoking_status}'])[0]
  
  prediction = model.predict([[gender, age, hypertension, heart_disease, ever_married,
       work_type, Residence_type, avg_glucose_level, bmi,
       smoking_status]])[0]
  print(prediction)
  if prediction == 0:
    return "Negative, you don't have Heart Stroke.  You are good to go"
  elif prediction == 1:
    return "Positive. You are most likely to have stroke. See a doctor NOW!"


def main():
  st.title('Heart Stroke Prediction')
  st.header('Fill in the below Features to determine wthether you have Stroke or not?')
  left_column, right_column = st.columns(2)
  with left_column:
    gender = st.radio(
        'Your gender?:',
        np.unique(df['gender']))
    

  age = st.number_input('Enter your Age?', value=0, step=1)

  hypertension = st.radio(
        'Have hypertension?:',
        np.unique(['Yes', 'No']))

  if hypertension == 'Yes':
    hypertension = 1
  elif hypertension == 'No':
    hypertension = 0


  heart_disease = st.radio(
        'Have Heart Disease?:',
        np.unique(['Yes', 'No']))
  if heart_disease == 'Yes':
    heart_disease = 1
  elif heart_disease == 'No':
    heart_disease = 0


  ever_married = st.radio(
        'Have you been ever married?:',
        np.unique(df['ever_married']))
  
  work_type = st.radio(
        'Work Type?:',
        np.unique(df['work_type']))

  residence_type = st.radio(
        'Residence Type?:',
        np.unique(df['Residence_type']))

  avg_glucose_level = st.slider('Average Glucose Level', 0.0, max(df["avg_glucose_level"]), 1.0)

  bmi = st.slider('BMI', 0.0, max(df["bmi"]), 1.0)

  smoking_status = st.radio(
        'Smoking Status?:',
        np.unique(df['smoking_status']))

  result = ''
  if st.button('Predict'):
    result = predict_stroke(gender, age, hypertension, heart_disease, ever_married,
       work_type, residence_type, avg_glucose_level, bmi,
       smoking_status)
  st.success(f'The output is: **{result}**')


if __name__ == '__main__':
  main()