app.py

import streamlit as st
import numpy as np
import joblib

# Load saved model + scaler
model = joblib.load(r"F:\Smart_Phone_ Addiction\smartphone_model.pkl")
scaler = joblib.load(r"F:\Smart_Phone_ Addiction\scaler.pkl")

st.set_page_config(page_title="Smartphone Addiction Predictor", layout="centered")

st.title("📱 Smartphone Addiction Predictor")

st.write("Enter user behavior details to predict addiction level.")

#  User Inputs 
age = st.slider("Age", 10, 60, 25)
gender = st.selectbox("Gender", ["Male", "Female"])

daily_screen_time = st.slider("Daily Screen Time (hours)", 0.0, 15.0, 5.0)
social_media = st.slider("Social Media Hours", 0.0, 10.0, 2.0)
gaming = st.slider("Gaming Hours", 0.0, 10.0, 1.0)
work_study = st.slider("Work/Study Hours", 0.0, 12.0, 4.0)
sleep = st.slider("Sleep Hours", 0.0, 12.0, 7.0)

notifications = st.slider("Notifications per Day", 0, 300, 50)
app_opens = st.slider("App Opens per Day", 0, 300, 40)
weekend_screen = st.slider("Weekend Screen Time", 0.0, 20.0, 6.0)

stress = st.selectbox("Stress Level", ["Low", "Medium", "High"])
impact = st.selectbox("Academic/Work Impact", ["Low", "Medium", "High"])
addiction_level = st.selectbox("Self-Reported Addiction Level", ["Low", "Medium", "High"])

#  Encoding 
def encode_inputs():
    gender_val = 1 if gender == "Male" else 0
    stress_val = {"Low": 0, "Medium": 1, "High": 2}[stress]
    impact_val = {"Low": 0, "Medium": 1, "High": 2}[impact]
    addiction_val = {"Low": 0, "Medium": 1, "High": 2}[addiction_level]

    total_usage = social_media + gaming + work_study
    sleep_deficit = 8 - sleep

    features = np.array([[
        age,
        gender_val,
        daily_screen_time,
        social_media,
        gaming,
        work_study,
        sleep,
        notifications,
        app_opens,
        weekend_screen,
        stress_val,
        impact_val,
        addiction_val,
        total_usage,
        sleep_deficit
    ]])

    return features

#  Prediction 
if st.button("Predict"):
    input_data = encode_inputs()
    input_scaled = scaler.transform(input_data)

    prediction = model.predict(input_scaled)[0]

    if prediction == 1:
        st.error(" High Risk of Smartphone Addiction")
    else:
        st.success(" Low Risk of Smartphone Addiction")