KomalDeep355/Gen-ZStressAnalyzer

1"""2Gen Z Social Media & Academic Performance Analytics — Streamlit Dashboard3Deployable on Hugging Face Spaces (SDK: streamlit).4"""5 6import sqlite37 8import matplotlib.pyplot as plt9import numpy as np10import pandas as pd11import seaborn as sns12import streamlit as st13from sklearn.ensemble import GradientBoostingClassifier, RandomForestRegressor14from sklearn.preprocessing import LabelEncoder15 16st.set_page_config(17    page_title="Gen Z Social Media & Academic Performance",18    page_icon="📱",19    layout="wide",20)21 22# ---------------------------------------------------------------------23# DATA LOADING24# ---------------------------------------------------------------------25@st.cache_data26def load_data():27    df = pd.read_csv("data/student_social_media_data.csv")28    return df29 30 31@st.cache_resource32def train_models(df):33    df_ml = df.copy()34    le_platform = LabelEncoder()35    le_year = LabelEncoder()36    le_gender = LabelEncoder()37    df_ml["platform_enc"] = le_platform.fit_transform(df_ml["platform_preference"])38    df_ml["year_enc"] = le_year.fit_transform(df_ml["year"])39    df_ml["gender_enc"] = le_gender.fit_transform(df_ml["gender"])40 41    features = ["daily_screen_time_hrs", "study_hours", "sleep_hours",42                "attendance_pct", "assignment_completion_pct",43                "platform_enc", "year_enc", "gender_enc"]44 45    X = df_ml[features]46    rf = RandomForestRegressor(n_estimators=300, max_depth=8, random_state=42)47    rf.fit(X, df_ml["cgpa"])48 49    gb = GradientBoostingClassifier(n_estimators=200, max_depth=3, random_state=42)50    gb.fit(X, df_ml["at_risk"])51 52    return rf, gb, le_platform, le_year, le_gender, features53 54 55df = load_data()56rf_model, gb_model, le_platform, le_year, le_gender, FEATURES = train_models(df)57 58# ---------------------------------------------------------------------59# SIDEBAR60# ---------------------------------------------------------------------61st.sidebar.title("📱 Filters")62platforms_sel = st.sidebar.multiselect(63    "Platform", sorted(df["platform_preference"].unique()),64    default=sorted(df["platform_preference"].unique())65)66years_sel = st.sidebar.multiselect(67    "Year", sorted(df["year"].unique()), default=sorted(df["year"].unique())68)69screen_range = st.sidebar.slider(70    "Daily Screen Time (hrs)",71    float(df["daily_screen_time_hrs"].min()),72    float(df["daily_screen_time_hrs"].max()),73    (float(df["daily_screen_time_hrs"].min()), float(df["daily_screen_time_hrs"].max()))74)75 76filtered = df[77    df["platform_preference"].isin(platforms_sel)78    & df["year"].isin(years_sel)79    & df["daily_screen_time_hrs"].between(*screen_range)80]81 82# ---------------------------------------------------------------------83# HEADER84# ---------------------------------------------------------------------85st.title("📱 Gen Z Social Media & Academic Performance Analytics")86st.caption("Does Screen Time Really Hurt Your CGPA? — Built on simulated student behavioral data.")87 88c1, c2, c3, c4 = st.columns(4)89c1.metric("Students (filtered)", len(filtered))90c2.metric("Avg CGPA", f"{filtered['cgpa'].mean():.2f}")91c3.metric("Avg Screen Time", f"{filtered['daily_screen_time_hrs'].mean():.1f} hrs/day")92c4.metric("Avg Exam Score", f"{filtered['exam_score'].mean():.1f}")93 94st.divider()95 96# ---------------------------------------------------------------------97# TABS98# ---------------------------------------------------------------------99tab1, tab2, tab3, tab4 = st.tabs(100    ["📊 EDA Dashboard", "🧮 SQL Explorer", "🤖 CGPA Predictor", "⚠️ Risk Classifier"]101)102 103with tab1:104    col1, col2 = st.columns(2)105 106    with col1:107        st.subheader("Screen Time Distribution")108        fig, ax = plt.subplots(figsize=(5, 3.2))109        sns.histplot(filtered["daily_screen_time_hrs"], bins=20, kde=True, ax=ax, color="#6C5CE7")110        st.pyplot(fig)111 112        st.subheader("Average CGPA by Platform")113        avg_plat = filtered.groupby("platform_preference")["cgpa"].mean().sort_values(ascending=False)114        st.bar_chart(avg_plat)115 116    with col2:117        st.subheader("Correlation Heatmap")118        num_cols = ["daily_screen_time_hrs", "study_hours", "sleep_hours",119                    "attendance_pct", "assignment_completion_pct", "exam_score", "cgpa"]120        fig2, ax2 = plt.subplots(figsize=(5, 4))121        sns.heatmap(filtered[num_cols].corr(), annot=True, fmt=".2f", cmap="coolwarm", ax=ax2)122        st.pyplot(fig2)123 124        st.subheader("CGPA by Screen Time Group")125        order = ["<2h", "2-4h", "4-6h", "6-8h", "8h+"]126        avg_group = (filtered.groupby("screen_time_group", observed=True)["cgpa"]127                     .mean().reindex(order))128        st.bar_chart(avg_group)129 130with tab2:131    st.subheader("Run SQL Queries Against the Dataset")132    st.caption("Table name: `students`")133    default_query = """SELECT platform_preference AS platform,134       ROUND(AVG(cgpa), 2) AS avg_cgpa,135       ROUND(AVG(exam_score), 1) AS avg_exam_score,136       COUNT(*) AS n_students137FROM students138GROUP BY platform_preference139ORDER BY avg_cgpa DESC;"""140    query = st.text_area("SQL query", value=default_query, height=160)141 142    if st.button("Run Query"):143        try:144            conn = sqlite3.connect(":memory:")145            filtered.to_sql("students", conn, index=False, if_exists="replace")146            result = pd.read_sql(query, conn)147            st.dataframe(result, use_container_width=True)148        except Exception as e:149            st.error(f"Query error: {e}")150 151with tab3:152    st.subheader("Predict CGPA from Behavioral Inputs")153    colA, colB, colC = st.columns(3)154    with colA:155        in_screen = st.slider("Daily screen time (hrs)", 0.0, 14.0, 5.5)156        in_study = st.slider("Study hours (hrs/day)", 0.0, 9.0, 3.0)157    with colB:158        in_sleep = st.slider("Sleep hours", 3.0, 10.0, 7.0)159        in_attendance = st.slider("Attendance %", 40.0, 100.0, 85.0)160    with colC:161        in_assignment = st.slider("Assignment completion %", 10.0, 100.0, 70.0)162        in_platform = st.selectbox("Platform preference", sorted(df["platform_preference"].unique()))163        in_year = st.selectbox("Year", sorted(df["year"].unique()))164        in_gender = st.selectbox("Gender", sorted(df["gender"].unique()))165 166    if st.button("Predict CGPA"):167        row = pd.DataFrame([{168            "daily_screen_time_hrs": in_screen,169            "study_hours": in_study,170            "sleep_hours": in_sleep,171            "attendance_pct": in_attendance,172            "assignment_completion_pct": in_assignment,173            "platform_enc": le_platform.transform([in_platform])[0],174            "year_enc": le_year.transform([in_year])[0],175            "gender_enc": le_gender.transform([in_gender])[0],176        }])[FEATURES]177        pred_cgpa = rf_model.predict(row)[0]178        st.success(f"Predicted CGPA: **{pred_cgpa:.2f}** / 10")179 180        importances = pd.Series(rf_model.feature_importances_, index=FEATURES).sort_values()181        st.caption("Model: Random Forest Regressor — feature importances below")182        st.bar_chart(importances)183 184with tab4:185    st.subheader("Academic Risk Classifier")186    st.caption("Flags students in the bottom ~20% CGPA band based on behavioral inputs.")187    colA, colB, colC = st.columns(3)188    with colA:189        r_screen = st.slider("Daily screen time (hrs) ", 0.0, 14.0, 7.0, key="r1")190        r_study = st.slider("Study hours (hrs/day) ", 0.0, 9.0, 1.5, key="r2")191    with colB:192        r_sleep = st.slider("Sleep hours ", 3.0, 10.0, 6.0, key="r3")193        r_attendance = st.slider("Attendance % ", 40.0, 100.0, 70.0, key="r4")194    with colC:195        r_assignment = st.slider("Assignment completion % ", 10.0, 100.0, 50.0, key="r5")196        r_platform = st.selectbox("Platform preference ", sorted(df["platform_preference"].unique()), key="r6")197        r_year = st.selectbox("Year ", sorted(df["year"].unique()), key="r7")198        r_gender = st.selectbox("Gender ", sorted(df["gender"].unique()), key="r8")199 200    if st.button("Assess Risk"):201        row = pd.DataFrame([{202            "daily_screen_time_hrs": r_screen,203            "study_hours": r_study,204            "sleep_hours": r_sleep,205            "attendance_pct": r_attendance,206            "assignment_completion_pct": r_assignment,207            "platform_enc": le_platform.transform([r_platform])[0],208            "year_enc": le_year.transform([r_year])[0],209            "gender_enc": le_gender.transform([r_gender])[0],210        }])[FEATURES]211        proba = gb_model.predict_proba(row)[0][1]212        label = "⚠️ At Risk" if proba >= 0.5 else "✅ On Track"213        st.metric("Risk Probability", f"{proba*100:.1f}%", label)214 215st.divider()216st.caption(217    "Note: This dataset is simulated for demonstration purposes. "218    "Relationships shown illustrate analytics/ML workflow, not verified real-world findings."219)