Code Editor for app.py

import streamlit as st
import pandas as pd
import numpy as np
import altair as alt

st.title("Missing Data Explorer")

# File uploader
uploaded_file = st.file_uploader("Upload a CSV file", type="csv")

if uploaded_file is not None:
    # Load data
    df = pd.read_csv(uploaded_file)
    
    # Calculate missing values
    missing_counts = df.isnull().sum()
    missing_percentages = (missing_counts / len(df) * 100).round(2)
    
    # Create missing values chart data
    missing_data = pd.DataFrame({
        'Column': missing_counts.index,
        'Missing Count': missing_counts.values,
        'Missing Percentage': missing_percentages.values
    })
    
    # Display missing values chart
    st.subheader("Missing Values Overview")
    
    chart = alt.Chart(missing_data).mark_bar().encode(
        x=alt.X('Missing Percentage:Q', title='Missing Percentage (%)'),
        y=alt.Y('Column:N', sort='-x', title='Column Name'),
        tooltip=['Column', 'Missing Count', 'Missing Percentage']
    ).properties(
        height=max(len(df.columns) * 25, 300)
    )
    
    st.altair_chart(chart, use_container_width=True)
    
    # Calculate missingness correlation
    st.subheader("Missingness Correlation Analysis")
    
    # Create binary matrix of missing values
    missing_binary = df.isnull().astype(int)
    
    # Calculate correlation between missing values
    missing_corr = missing_binary.corr()
    
    # For each variable, show the top 5 correlated variables in terms of missingness
    for col in df.columns:
        if missing_counts[col] > 0:  # Only show for columns with missing values
            st.write(f"\n**Top 5 missingness correlations for {col}:**")
            
            # Get correlations for this column
            col_corr = missing_corr[col].sort_values(ascending=False)
            # Remove self-correlation and get top 5
            top_5_corr = col_corr[col_corr.index != col][:5]
            
            # Create correlation data for this column
            corr_data = pd.DataFrame({
                'Related Column': top_5_corr.index,
                'Correlation': top_5_corr.values.round(3)
            })
            
            st.table(corr_data)
else:
    st.write("Please upload a CSV file to begin the analysis.")