app.py

# -*- coding: utf-8 -*-
"""
Created on Tue Feb 11 14:11:12 2025

@author: Newdeskarl
"""

import streamlit as st
import pandas as pd
import plotly.express as px
import plotly.graph_objects as go

# Set page config
st.set_page_config(
    page_title="Cluster Analysis Dashboard",
    page_icon="📊",
    layout="wide"
)

# Add title
st.title("Cluster Analysis Dashboard")

# Load data
@st.cache_data
def load_data():
    df = pd.read_excel("pi_gouv_avec_cluster.xlsx")
    return df

try:
    df = load_data()
    
    # Create columns for filters
    col1, col2, col3 = st.columns(3)
    
    # Filter columns to exclude certain fields
    exclude_cols = ['NomPI', 'Lat', 'Long', 'kmeans_clusters']
    columns = [col for col in df.columns if col not in exclude_cols]
    
    with col1:
        selected_column = st.selectbox(
            "Select Column to Analyze:",
            columns
        )
    
    with col2:
        selected_pi = st.selectbox(
            "Select NomPI:",
            ["All"] + list(df['NomPI'].unique())
        )
    
    with col3:
        selected_cluster = st.selectbox(
            "Select Cluster:",
            ["All"] + list(df['kmeans_clusters'].unique().astype(str))
        )
    
    # Filter data based on selections
    filtered_df = df.copy()
    
    if selected_pi != "All":
        filtered_df = filtered_df[filtered_df['NomPI'] == selected_pi]
        
    if selected_cluster != "All":
        filtered_df = filtered_df[filtered_df['kmeans_clusters'] == int(selected_cluster)]
    
    # Create pie chart
    value_counts = filtered_df[selected_column].value_counts()
    
    fig = go.Figure(data=[go.Pie(
        labels=value_counts.index,
        values=value_counts.values,
        hole=0.3,
        textinfo='label+percent',
        hovertemplate="<b>%{label}</b><br>" +
                      "Count: %{value}<br>" +
                      "Percentage: %{percent}<extra></extra>"
    )])
    
    # Update layout
    title = f"Distribution of {selected_column}"
    if selected_pi != "All":
        title += f" for {selected_pi}"
    if selected_cluster != "All":
        title += f" in Cluster {selected_cluster}"
        
    fig.update_layout(
        title=title,
        showlegend=True,
        height=600
    )
    
    # Display the chart
    st.plotly_chart(fig, use_container_width=True)
    
    # Display detailed statistics
    st.subheader("Detailed Statistics")
    
    # Create a DataFrame with the statistics
    stats_df = pd.DataFrame({
        'Value': value_counts.index,
        'Count': value_counts.values,
        'Percentage': (value_counts.values / value_counts.sum() * 100).round(2)
    })
    
    # Display the statistics in a nice format
    st.dataframe(
        stats_df.style.format({'Percentage': '{:.2f}%'}),
        hide_index=True
    )

except Exception as e:
    st.error(f"An error occurred: {str(e)}")
    st.info("Please make sure the file 'pi_gouv_avec_cluster.xlsx' is in the same directory as this script.")

# Add footer with instructions
st.markdown("---")
st.markdown("""
### How to use this dashboard:
1. Select a column to analyze from the first dropdown
2. Optionally filter by specific NomPI
3. Optionally filter by specific cluster
4. The pie chart and statistics will update automatically
""")