app.py

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

# Load data function
def load_data(uploaded_file):
    if uploaded_file is not None:
        df = pd.read_csv(uploaded_file)
        df.fillna(0, inplace=True)
        if '出表日期' in df.columns:
            df['出表日期'] = df['出表日期'].astype(str)
        if '公司代號' in df.columns:
            df['公司代號'] = df['公司代號'].astype(str)
        return df
    else:
        st.warning("請上傳檔案。")
        return None

# Merge dataframes
def merge_dataframes(df1, df2, on_columns):
    if df1 is None or df2 is None:
        return None
    for col in on_columns:
        if col in df1.columns and col in df2.columns:
            df1[col] = df1[col].astype(str)
            df2[col] = df2[col].astype(str)
    return pd.merge(df1, df2, on=on_columns, how="outer")

# Filter dataframe
def filter_dataframe(df, prefix):
    return df[df['公司代號'].astype(str).str.startswith(prefix)]

# Get specific company data
def get_specific_company(df, company_code):
    return df[df['公司代號'] == company_code]

# Plot radar chart
def plot_radar_chart(avg_values, specific_company_values, categories, prefix, specific_company_name):
    fig = go.Figure()

    fig.add_trace(go.Scatterpolar(
        r=avg_values,
        theta=categories,
        fill='toself',
        name=f"股號前兩位『{prefix}』的族群"
    ))

    fig.add_trace(go.Scatterpolar(
        r=specific_company_values,
        theta=categories,
        fill='toself',
        name=f'{specific_company_name}'
    ))

    fig.update_layout(
        polar=dict(radialaxis=dict(visible=True, range=[0, 100])),
        showlegend=True,
        title="董事會和投資人溝通指標比較"
    )

    st.plotly_chart(fig)

# Plot emission chart
def plot_emission_chart(filtered_df, avg_emissions, prefix):
    emission_columns = ['範疇一排放量(噸CO2e)', '範疇二排放量(噸CO2e)', '範疇三排放量(噸CO2e)']
    fig = go.Figure()

    for scope, color in zip(emission_columns, ['blue', 'green', 'red']):
        fig.add_trace(go.Bar(
            x=filtered_df['公司名稱'],
            y=filtered_df[scope],
            name=scope,
            marker_color=color
        ))

        fig.add_trace(go.Scatter(
            x=filtered_df['公司名稱'],
            y=[avg_emissions[scope]] * len(filtered_df),
            mode='lines',
            line=dict(color=color, dash='dash'),
            name=f'{scope}平均值'
        ))

    fig.update_layout(
        title=f"代號前兩位『{prefix}』的族群 - 各範疇排放量",
        barmode='group',
        xaxis_title="公司名稱",
        yaxis_title="排放量(噸CO2e)"
    )

    st.plotly_chart(fig)

# Plot energy usage
def plot_energy_usage(filtered_df, avg_energy_usage):
    fig_energy = px.bar(filtered_df, x='公司名稱', y='使用率(再生能源/總能源)', title="再生能源使用率")

    fig_energy.add_trace(go.Scatter(
        x=filtered_df['公司名稱'],
        y=[avg_energy_usage] * len(filtered_df),
        mode='lines',
        line=dict(color='red', dash='dash'),
        name='群體平均值'
    ))

    fig_energy.update_layout(
        yaxis_title="再生能源使用率 (%)",
        xaxis_title="公司名稱"
    )

    st.plotly_chart(fig_energy)
# Plot waste management box plots
def plot_waste_management(df_group, df_specific, company_name):
    columns_to_analyze = ['有害廢棄物量-數據(公噸)', '非有害廢棄物量-數據(公噸)', '總重量(有害+非有害)-數據(公噸)', '廢棄物密集度-密集度(公噸/單位)']

    fig = go.Figure()

    # Loop through the columns and plot box plot for each column
    for col in columns_to_analyze:
        fig.add_trace(go.Box(y=df_group[col], name=f'母群體-{col}', boxmean=True, boxpoints='outliers'))

        if not df_specific.empty:
            specific_value = df_specific[col].values[0]
            
            # Highlight specific company's value
            fig.add_trace(go.Scatter(
                y=[specific_value],
                x=[f'母群體-{col}'],
                mode='markers',
                name=f'{company_name}-{col}',
                marker=dict(color='red', size=11, symbol='star'),
                showlegend=True,
                hovertext=f'公司名稱: {company_name}, 值: {specific_value}'
            ))

    # Update layout
    fig.update_layout(
        title=f"廢棄物統計數據箱型圖 (包含指定公司名稱 {company_name} 數據)",
        yaxis_title="數值 (公噸)",
        xaxis_title="廢棄物項目",
        boxmode='group'
    )

    # Display the plot in Streamlit
    st.plotly_chart(fig)

    
# Main function update
def main():
    st.title("ESG analytic dashboard")
    
    # File upload
    st.sidebar.header("上傳CSV檔案- https://reurl.cc/yvAEql")
    investor_file = st.sidebar.file_uploader("上傳【投資人溝通】.csv", type=["csv"])
    board_file = st.sidebar.file_uploader("上傳【董事會】.csv", type=["csv"])
    emission_file = st.sidebar.file_uploader("上傳【溫室氣體排放】.csv", type=["csv"])
    energy_file = st.sidebar.file_uploader("上傳【能源管理】.csv", type=["csv"])
    waste_file = st.sidebar.file_uploader("上傳【廢棄物管理】.csv", type=["csv"])

    # Load data
    investor_df = load_data(investor_file)
    board_df = load_data(board_file)
    emission_df = load_data(emission_file)
    energy_df = load_data(energy_file)
    waste_df = load_data(waste_file)

    # Merge data
    merged_df1 = merge_dataframes(investor_df, board_df, ["公司代號", "公司名稱", "出表日期", "報告年度"])
    merged_df2 = merge_dataframes(emission_df, energy_df, ["公司代號", "公司名稱", "出表日期", "報告年度"])

    # User input
    prefix = st.sidebar.text_input("輸入公司代號前兩位")
    specific_company_code = st.sidebar.text_input("輸入四位數字公司代號")

    # Waste management analysis
    if waste_df is not None and prefix:
        waste_df['公司代號前兩位'] = waste_df['公司代號'].astype(str).str[:2]
        df_group = waste_df[waste_df['公司代號前兩位'] == prefix]
        df_specific = waste_df[waste_df['公司代號'] == specific_company_code]

        if not df_specific.empty:
            company_name = df_specific['公司名稱'].values[0]
            plot_waste_management(df_group, df_specific, company_name)
        else:
            st.warning(f"找不到公司代號為 {specific_company_code} 的廢棄物管理數據")



    # Handle 投資人溝通和董事會資料
    if merged_df1 is not None and prefix and specific_company_code:
        columns_of_interest = ['董事出席董事會出席率', '董事進修時數符合進修要點比率', '公司年度召開法說會次數(次)']
        for col in ['董事出席董事會出席率', '董事進修時數符合進修要點比率']:
            merged_df1[col] = merged_df1[col].replace({'%': ''}, regex=True).astype(float)

        filtered_df1 = filter_dataframe(merged_df1, prefix)
        avg_values = filtered_df1[columns_of_interest].mean()
        specific_company_df1 = get_specific_company(merged_df1, specific_company_code)

        if not specific_company_df1.empty:
            specific_company_name = specific_company_df1['公司名稱'].iloc[0]
            specific_company_values = specific_company_df1[columns_of_interest].iloc[0]
            plot_radar_chart(avg_values, specific_company_values, ['董事出席率', '董事進修時數符合比率', '年度法說會次數'], prefix, specific_company_name)
        else:
            st.warning(f"找不到公司代號 {specific_company_code} 的資料")

    # Handle 溫室氣體排放和能源管理資料
    if merged_df2 is not None and prefix:
        emission_columns = ['範疇一排放量(噸CO2e)', '範疇二排放量(噸CO2e)', '範疇三排放量(噸CO2e)']
        energy_column = '使用率(再生能源/總能源)'
        merged_df2[energy_column] = merged_df2[energy_column].replace({'%': ''}, regex=True).astype(float)

        filtered_df2 = filter_dataframe(merged_df2, prefix)
        specific_company_df2 = get_specific_company(merged_df2, specific_company_code)

        if not filtered_df2.empty:
            avg_emissions = filtered_df2[emission_columns].mean()
            plot_emission_chart(filtered_df2, avg_emissions, prefix)

            avg_energy_usage = filtered_df2[energy_column].mean()
            plot_energy_usage(filtered_df2, avg_energy_usage)

            if not specific_company_df2.empty:
                specific_energy_usage = specific_company_df2[energy_column].iloc[0]
                comparison_data = {
                    '公司名稱': [specific_company_df2['公司名稱'].iloc[0], f"{prefix} 母群體平均"],
                    '再生能源使用率 (%)': [specific_energy_usage, avg_energy_usage]
                }
                comparison_df = pd.DataFrame(comparison_data)
                st.write("\n再生能源使用率比較表格：")
                st.write(comparison_df)
            else:
                st.warning(f"找不到公司代號 {specific_company_code} 的能源管理數據")
        else:
            st.warning(f"找不到前兩碼為 {prefix} 的公司數據")

if __name__ == "__main__":
    main()

# https://drive.google.com/drive/folders/1uZnryIluMn-bszuHsFMbLecbL6Vq3HyI?usp=sharing