app.py

import streamlit as st
import pandas as pd
import tensorflow as tf
from tensorflow.keras.models import load_model
from sklearn.preprocessing import StandardScaler
import numpy as np
import pickle

# Load the model
model = load_model('my_model.h5')

# Load the scaler
with open('scaler.pkl', 'rb') as f:
    scaler = pickle.load(f)

# Create a sample input dataframe
sample_data = {
    'Application mode': [1],
    'Application order': [1],
    'Previous qualification (grade)': [125],
    'Admission grade': [119],
    'Displaced': [1],
    'Debtor': [0],
    'Tuition fees up to date': [1],
    'Gender': [0],
    'Scholarship holder': [0],
    'Age at enrollment': [18],
    'Curricular units 1st sem (enrolled)': [6],
    'Curricular units 1st sem (evaluations)': [8],
    'Curricular units 1st sem (approved)': [4],
    'Curricular units 1st sem (grade)': [11],
    'Curricular units 2nd sem (enrolled)': [6],
    'Curricular units 2nd sem (evaluations)': [9],
    'Curricular units 2nd sem (approved)': [0],
    'Curricular units 2nd sem (grade)': [0],
    'Curricular units 2nd sem (without evaluations)': [1]
}

sample_df = pd.DataFrame(sample_data)

# Function to get user input
def get_user_input():
    input_data = {
        'Application mode': 1,
        'Application order': 1,
        'Previous qualification (grade)': st.number_input('Previous qualification (grade)', value=int(sample_df['Previous qualification (grade)'][0])),
        'Admission grade': st.number_input('Admission grade', value=int(sample_df['Admission grade'][0])),
        'Displaced': st.selectbox('Displaced', options=[0, 1], format_func=lambda x: 'Yes' if x == 1 else 'No'),
        'Debtor': st.selectbox('Debtor', options=[0, 1], format_func=lambda x: 'Yes' if x == 1 else 'No'),
        'Tuition fees up to date': st.selectbox('Tuition fees up to date', options=[0, 1], format_func=lambda x: 'Yes' if x == 1 else 'No'),
        'Gender': st.selectbox('Gender', options=[0, 1], format_func=lambda x: 'Female' if x == 1 else 'Male'),
        'Scholarship holder': st.selectbox('Scholarship holder', options=[0, 1], format_func=lambda x: 'Yes' if x == 1 else 'No'),
        'Age at enrollment': st.number_input('Age at enrollment', value=int(sample_df['Age at enrollment'][0])),
        'Curricular units 1st sem (enrolled)': st.number_input('Curricular units 1st sem (enrolled)', value=int(sample_df['Curricular units 1st sem (enrolled)'][0])),
        'Curricular units 1st sem (evaluations)': st.number_input('Curricular units 1st sem (evaluations)', value=int(sample_df['Curricular units 1st sem (evaluations)'][0])),
        'Curricular units 1st sem (approved)': st.number_input('Curricular units 1st sem (approved)', value=int(sample_df['Curricular units 1st sem (approved)'][0])),
        'Curricular units 1st sem (grade)': st.number_input('Curricular units 1st sem (grade)', value=int(sample_df['Curricular units 1st sem (grade)'][0])),
        'Curricular units 2nd sem (enrolled)': st.number_input('Curricular units 2nd sem (enrolled)', value=int(sample_df['Curricular units 2nd sem (enrolled)'][0])),
        'Curricular units 2nd sem (evaluations)': st.number_input('Curricular units 2nd sem (evaluations)', value=int(sample_df['Curricular units 2nd sem (evaluations)'][0])),
        'Curricular units 2nd sem (approved)': st.number_input('Curricular units 2nd sem (approved)', value=int(sample_df['Curricular units 2nd sem (approved)'][0])),
        'Curricular units 2nd sem (grade)': st.number_input('Curricular units 2nd sem (grade)', value=int(sample_df['Curricular units 2nd sem (grade)'][0])),
        'Curricular units 2nd sem (without evaluations)': st.number_input('Curricular units 2nd sem (without evaluations)', value=int(sample_df['Curricular units 2nd sem (without evaluations)'][0]))
    }
    return pd.DataFrame(input_data, index=[0])

# Streamlit app
st.title('Student Outcome Prediction 🧑‍🎓')
st.write('This app predicts whether a student is enrolled, graduated or dropout. 🎓')
st.write('Model is trained on a dataset by UC Irvine, acquired from several disjoint higher education institution databases. It has 82% accuracy. 🏫')

# Get user input
user_input_df = get_user_input()

# Normalize the user input
user_input_scaled = scaler.transform(user_input_df)

# Make prediction
prediction = model.predict(user_input_scaled)
predicted_class = prediction.argmax(axis=1)[0]

# Map the predicted class back to words
target_mapping = {0: 'Graduate', 1: 'Dropout', 2: 'Enrolled'}
predicted_label = target_mapping[predicted_class]

# Display the prediction
st.header(f'The predicted outcome is: ***{predicted_label}***')
st.image('https://d13b2ieg84qqce.cloudfront.net/c42fed0bb5d4f458b8606152e7dec885cd3e751d')