project_1.py

import pandas as pd
import streamlit as st
import plotly.express as px
import numpy as np
import scipy.stats as stats

# Elegant and Centered Title
st.markdown(
    """
    <h1 style='text-align: center; color: red;'>Tips Dataset Explorer</h1>
    """,
    unsafe_allow_html=True
)

# Upload Section
file = st.file_uploader("📂 Upload a CSV file", type=["csv"])

if file is not None:
    df = pd.read_csv(file)
    st.markdown("### 🧾 Preview of the Dataset:")
    st.write(df.head())

    categorical_cols = df.select_dtypes(include='object').columns.tolist()
    numerical_cols = df.select_dtypes(include=['int64', 'float64']).columns.tolist()

    st.markdown("## 🔍 Univariate Analysis")

    # ========================== #
    #      CATEGORICAL PART      #
    # ========================== #
    st.markdown("### 🟨 Categorical Feature Analysis")

    selected_cat = st.selectbox("Select a categorical feature", categorical_cols, key='cat_feature')

    if selected_cat:
        col_data = df[selected_cat]
        st.write(f"**Feature: `{selected_cat}`**")
        st.write(f"🔢 Unique values: {col_data.unique()}")
        st.write(f"🧮 Number of unique values: {col_data.nunique()}")
        st.write(f"📌 Mode: {col_data.mode()[0]}")
        st.write(f"❓ Missing values: {col_data.isnull().sum()}")

        value_counts = col_data.value_counts()
        proportion = (value_counts / len(df)).round(3) * 100

        st.write("📊 Frequency Table:")
        freq_table = pd.DataFrame({ 'Count': value_counts, 'Percentage (%)': proportion })
        st.write(freq_table)

        # Entropy
        entropy = stats.entropy(value_counts)
        st.write(f"🧠 Entropy (uncertainty in distribution): `{entropy:.3f}`")

        # Charts
        st.write("📊 Bar Chart:")
        st.bar_chart(value_counts)

        if st.checkbox("📈 Show Pie Chart", key="pie_cat"):
            fig_pie = px.pie(names=value_counts.index, values=value_counts.values, 
                             title=f'Distribution of {selected_cat}')
            st.plotly_chart(fig_pie)

        # Cross-tab Heatmap
        other_cat = st.selectbox("Compare with another categorical feature", 
                                 [col for col in categorical_cols if col != selected_cat], key='cat_cross')
        if other_cat:
            cross_tab = pd.crosstab(df[selected_cat], df[other_cat])
            st.markdown(f"### 🔁 Cross-tabulation: `{selected_cat}` vs `{other_cat}`")
            st.dataframe(cross_tab.style.background_gradient(cmap='viridis'))

    # ========================== #
    #      NUMERICAL PART        #
    # ========================== #
    st.markdown("### 🟦 Numerical Feature Analysis")

    selected_num = st.selectbox("Select a numerical feature", numerical_cols, key='num_feature')

    if selected_num:
        col_data = df[selected_num].dropna()

        st.write(f"**Feature: `{selected_num}`**")
        st.write(f"📏 Mean: {col_data.mean():.2f}")
        st.write(f"🪙 Median: {col_data.median():.2f}")
        st.write(f"📉 Min: {col_data.min()} | 📈 Max: {col_data.max()}")
        st.write(f"📊 Std: {col_data.std():.2f} | 📐 Variance: {col_data.var():.2f}")
        st.write(f"📈 Skewness: {col_data.skew():.2f} | 🔺 Kurtosis: {col_data.kurt():.2f}")
        st.write(f"❓ Missing Values: {df[selected_num].isnull().sum()}")

        # Quartiles
        Q1 = col_data.quantile(0.25)
        Q2 = col_data.quantile(0.50)
        Q3 = col_data.quantile(0.75)
        IQR = Q3 - Q1
        st.write(f"🔢 Quartiles: Q1={Q1:.2f}, Q2={Q2:.2f}, Q3={Q3:.2f}, IQR={IQR:.2f}")

        # Outliers based on Z-score
        z_scores = np.abs(stats.zscore(col_data))
        outliers = (z_scores > 3).sum()
        st.write(f"🚨 Outliers (Z-score > 3): {outliers}")

        # Distribution Plot
        st.write("📊 Distribution (Histogram):")
        fig_hist = px.histogram(df, x=selected_num, nbins=30, title=f'Distribution of {selected_num}')
        st.plotly_chart(fig_hist)

        # Box Plot
        st.write("📦 Box Plot:")
        fig_box = px.box(df, y=selected_num, title=f'Box Plot of {selected_num}')
        st.plotly_chart(fig_box)

        # KDE Plot Option
        if st.checkbox("📈 Show KDE (Density Estimation)", key='kde'):
            fig_kde = px.histogram(df, x=selected_num, nbins=30, marginal="violin", title=f'KDE of {selected_num}')
            st.plotly_chart(fig_kde)

        # Binning
        if st.checkbox("🧊 Show Value Bins", key='bins'):
            bins = st.slider("Number of bins", min_value=2, max_value=20, value=5)
            binned_col = pd.cut(col_data, bins=bins)
            bin_counts = binned_col.value_counts().sort_index()
            st.write(bin_counts)

            fig_bin_pie = px.pie(names=bin_counts.index.astype(str), values=bin_counts.values,
                                 title=f'Binned Distribution of {selected_num}')
            st.plotly_chart(fig_bin_pie)

        # Log Transformation Preview
        if st.checkbox("🔄 Preview Log Transformation", key='log_transform'):
            df['log_' + selected_num] = np.log1p(col_data)
            fig_log = px.histogram(df, x='log_' + selected_num, nbins=30, 
                                   title=f'Log-Transformed Histogram of {selected_num}')
            st.plotly_chart(fig_log)

    # Summary Table
    st.markdown("### 📋 Summary Statistics")
    st.dataframe(df[numerical_cols].describe().T.style.background_gradient(cmap='Blues'))

else:
    st.warning("📁 Please upload a CSV file to begin analysis.")