Update app.py

app.py

@@ -9,24 +9,16 @@ warnings.filterwarnings('ignore')
 st.set_page_config(
     page_title="SAP Sales KPI Dashboard",
     page_icon="📊",
-    layout="wide",
-    initial_sidebar_state="expanded"
+    layout="wide"
 )
 
 # --- Custom CSS ---
 st.markdown("""
 <style>
-    .main-header {
-        font-size: 2.8rem; font-weight: bold; color: #1f4e79;
-        text-align: center; margin-bottom: 1rem;
-    }
-    .kpi-card {
-        background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
-        padding: 1.5rem; border-radius: 15px; color: white;
-        margin: 0.5rem 0; box-shadow: 0 4px 6px rgba(0,0,0,0.1);
-    }
-    .kpi-value { font-size: 2.5rem; font-weight: bold; margin-bottom: 0.5rem; }
-    .kpi-label { font-size: 1rem; opacity: 0.9; }
+    .main-header { font-size: 2.8rem; font-weight: bold; color: #1f4e79; text-align: center; }
+    .kpi-card { background: #f8f9fa; padding: 1.5rem; border-radius: 10px; border-left: 5px solid #667eea; margin-bottom: 1rem; }
+    .kpi-value { font-size: 2.5rem; font-weight: bold; color: #1f4e79; }
+    .kpi-label { font-size: 1rem; color: #555; }
 </style>
 """, unsafe_allow_html=True)
 
@@ -34,185 +26,139 @@ st.markdown("""
 @st.cache_data(ttl=3600)
 def load_kaggle_sap_data():
     try:
+        # Check for secrets
         if 'KAGGLE_USERNAME' not in st.secrets or 'KAGGLE_KEY' not in st.secrets:
-            st.error("Kaggle credentials not found in Streamlit secrets.")
-            return None
+            return "Kaggle credentials not found in Streamlit secrets."
 
         os.environ['KAGGLE_USERNAME'] = st.secrets['KAGGLE_USERNAME']
         os.environ['KAGGLE_KEY'] = st.secrets['KAGGLE_KEY']
-
+        
         import kaggle
-
         dataset_name = "mustafakeser4/sap-dataset-bigquery-dataset"
         download_path = "./kaggle_data"
 
+        # Download only if files don't exist
         if not os.path.exists(os.path.join(download_path, 'vbak.csv')):
-            kaggle.api.authenticate()
-            kaggle.api.dataset_download_files(dataset_name, path=download_path, unzip=True)
-            st.success("Successfully downloaded dataset from Kaggle.")
-        else:
-            st.info("Using cached Kaggle dataset.")
+            with st.spinner("Downloading dataset from Kaggle... This may take a moment."):
+                kaggle.api.authenticate()
+                kaggle.api.dataset_download_files(dataset_name, path=download_path, unzip=True)
 
+        # Load tables
         tables = {}
-        required_files = {
-            'vbak': 'vbak.csv', 'vbap': 'vbap.csv',
-            'kna1': 'kna1.csv', 'makt': 'makt.csv'
-        }
-        for name, filename in required_files.items():
+        for name, filename in {'vbak': 'vbak.csv', 'vbap': 'vbap.csv', 'kna1': 'kna1.csv', 'makt': 'makt.csv'}.items():
             file_path = os.path.join(download_path, filename)
             if os.path.exists(file_path):
                 tables[name] = pd.read_csv(file_path, low_memory=False)
-            else:
-                tables[name] = pd.DataFrame()
-        
         return tables
 
     except Exception as e:
-        st.error(f"Error during Kaggle data loading: {e}")
-        return None
+        return f"Error during Kaggle data loading: {e}"
 
 # --- Data Processing & Analytics ---
 @st.cache_data
 def create_sales_analytics(_tables):
-    if not _tables: return None
-
     try:
         vbak = _tables['vbak'].copy()
         vbap = _tables['vbap'].copy()
         kna1 = _tables['kna1'].copy()
         makt = _tables['makt'].copy()
 
+        # Normalize column names
         for df in [vbak, vbap, kna1, makt]:
             df.columns = [col.upper() for col in df.columns]
 
         makt_en = makt[makt['SPRAS'] == 'E']
 
-        sales_data = pd.merge(
-            vbak[['VBELN', 'KUNNR', 'VKORG', 'VTWEG', 'ERDAT']],
-            vbap[['VBELN', 'MATNR', 'NETWR']],
-            on='VBELN', how='inner'
-        )
-        sales_data = sales_data.merge(kna1[['KUNNR', 'NAME1', 'LAND1']], on='KUNNR', how='left')
-        sales_data = sales_data.merge(makt_en[['MATNR', 'MAKTX']], on='MATNR', how='left')
-
+        # Merge and process data
+        sales_data = pd.merge(vbak, vbap, on='VBELN', how='inner')
+        sales_data = sales_data.merge(kna1, on='KUNNR', how='left')
+        sales_data = sales_data.merge(makt_en, on='MATNR', how='left')
+        
         sales_data['NETWR'] = pd.to_numeric(sales_data['NETWR'], errors='coerce').fillna(0)
         sales_data['ERDAT'] = pd.to_datetime(sales_data['ERDAT'], errors='coerce')
         
-        # Limit the dataset size after merging for better performance
-        return sales_data.head(20000)
+        return sales_data
 
     except Exception as e:
-        st.error(f"Error processing sales data: {e}")
-        return None
+        return f"Error processing sales data: {e}"
 
 # --- UI Components ---
-def create_kpi_card(title, value, format_type="currency"):
-    if format_type == "currency": value_str = f"${value:,.0f}"
-    elif format_type == "number": value_str = f"{value:,.0f}"
-    else: value_str = str(value)
-    
+def create_kpi_card(title, value, format_str="${:,.0f}"):
     st.markdown(f"""
     <div class="kpi-card">
-        <div class="kpi-value">{value_str}</div>
+        <div class="kpi-value">{value:{format_str.split('{')[1].split('}')[0]}}</div>
         <div class="kpi-label">{title}</div>
     </div>
     """, unsafe_allow_html=True)
 
-def main():
-    # --- Header ---
-    st.markdown('<h1 class="main-header">🎯 SAP Sales KPI Dashboard</h1>', unsafe_allow_html=True)
-    st.markdown("""
-    <div style="text-align: center; margin-bottom: 2rem;">
-        <p style="font-size: 1.2rem; color: #666;">
-        Real SAP ERP Sales Data from Kaggle | Customer • Regional • Channel • Product KPIs
-        </p>
-    </div>
-    """, unsafe_allow_html=True)
+# --- Main App Logic ---
+st.markdown('<h1 class="main-header">🎯 SAP Sales KPI Dashboard</h1>', unsafe_allow_html=True)
 
-    # --- Setup Instructions ---
-    with st.expander("🔑 Kaggle API Setup Instructions", expanded=True):
-        st.info("This app requires Kaggle credentials. If it fails to load, please add your credentials to the secrets.")
-        st.markdown("""
-        1.  Go to **Kaggle > Account > API > Create New Token** to download `kaggle.json`.
-        2.  In your Hugging Face Space, go to **Settings > Secrets**.
-        3.  Add two secrets: `KAGGLE_USERNAME` and `KAGGLE_KEY` with the values from your file.
-        4.  Restart the Space.
-        """)
-
-    # --- Data Loading and Processing ---
-    tables = load_kaggle_sap_data()
-    if not tables: return
-
-    sales_df = create_sales_analytics(tables)
-    if sales_df is None or sales_df.empty: return
-
-    st.success(f"✅ Loaded and processed {len(sales_df):,} real SAP sales records!")
-
-    # --- Main Dashboard ---
-    st.sidebar.header("📊 Real SAP Dataset")
-    st.sidebar.success(f"**Analysis Records:** {len(sales_df):,}\n\n**Data Source:** Kaggle")
-
-    # --- KPIs ---
-    st.subheader("Sales KPIs from Real SAP Data")
-    col1, col2, col3, col4 = st.columns(4)
-    with col1: create_kpi_card("Total Revenue", sales_df['NETWR'].sum())
-    with col2: create_kpi_card("Active Customers", sales_df['KUNNR'].nunique(), "number")
-    with col3: create_kpi_card("Avg Order Value", sales_df[sales_df['NETWR'] > 0]['NETWR'].mean())
-    with col4: create_kpi_card("Sales Orders", sales_df['VBELN'].nunique(), "number")
-
-    # --- Analytics Tabs ---
-    tab1, tab2, tab3, tab4 = st.tabs(["👥 Top Customers", "🌍 Regional Analysis", "📈 Distribution Channels", "🛍️ Top Products"])
-
-    # Customer Analysis
-    with tab1:
-        st.subheader("Top 10 Customers by Revenue")
-        customer_summary = sales_df.groupby('NAME1')['NETWR'].sum().reset_index()
-        top_customers = customer_summary.nlargest(10, 'NETWR')
-        fig = px.bar(top_customers, x='NETWR', y='NAME1', orientation='h', title="Top 10 Customers", labels={'NETWR': 'Revenue ($)', 'NAME1': 'Customer'}, color='NETWR', color_continuous_scale='Blues')
-        st.plotly_chart(fig.update_layout(yaxis={'categoryorder': 'total ascending'}), use_container_width=True)
-
-    # Regional Analysis
-    with tab2:
-        st.subheader("Revenue by Country")
-        # **FIXED: Create a clean summary DataFrame for the pie chart**
-        regional_summary = sales_df.groupby('LAND1')['NETWR'].sum().reset_index()
-        top_regions = regional_summary.nlargest(10, 'NETWR')
-        
-        fig = px.pie(top_regions, 
-                     values='NETWR', 
-                     names='LAND1',
-                     title="Top 10 Countries by Revenue")
-        st.plotly_chart(fig, use_container_width=True)
-
-    # Distribution Channel Analysis
-    with tab3:
-        st.subheader("Revenue by Distribution Channel")
-        # **FIXED: Create a clean summary DataFrame for the bar chart**
-        channel_summary = sales_df.groupby('VTWEG')['NETWR'].sum().reset_index()
-        
-        # Rename channel codes for clarity if they are numeric
-        channel_summary['VTWEG'] = channel_summary['VTWEG'].astype(str)
-
-        fig = px.bar(channel_summary, 
-                     x='VTWEG', 
-                     y='NETWR',
-                     title="Total Revenue by Distribution Channel", 
-                     labels={'NETWR': 'Total Revenue ($)', 'VTWEG': 'Distribution Channel'}, 
-                     color='NETWR', 
-                     color_continuous_scale='Plasma')
-        st.plotly_chart(fig, use_container_width=True)
-
-    # Product Analysis
-    with tab4:
-        st.subheader("Top 10 Products by Revenue")
-        product_summary = sales_df.groupby('MAKTX')['NETWR'].sum().reset_index()
-        top_products = product_summary.nlargest(10, 'NETWR')
-        fig = px.bar(top_products, x='NETWR', y='MAKTX', orientation='h', title="Top 10 Products", labels={'NETWR': 'Revenue ($)', 'MAKTX': 'Product'}, color='NETWR', color_continuous_scale='Greens')
-        st.plotly_chart(fig.update_layout(yaxis={'categoryorder': 'total ascending'}), use_container_width=True)
-
-    # --- Footer ---
-    st.markdown("---")
-    st.markdown("<p style='text-align: center;'>Built with Streamlit • 100% Real SAP ERP Data from Kaggle</p>", unsafe_allow_html=True)
-
-if __name__ == "__main__":
-    main()
+# Cache clearing button
+if st.sidebar.button("🔄 Clear Cache & Rerun"):
+    st.cache_data.clear()
+    st.rerun()
+
+st.sidebar.title("Dashboard Controls")
+
+# Load data and handle errors
+raw_tables = load_kaggle_sap_data()
+if isinstance(raw_tables, str):
+    st.error(raw_tables)
+    st.stop()
+
+sales_df = create_sales_analytics(raw_tables)
+if isinstance(sales_df, str):
+    st.error(sales_df)
+    st.stop()
+
+st.success(f"✅ Loaded and processed {len(sales_df):,} real SAP sales records!")
+
+# --- Sidebar Filters ---
+st.sidebar.header("Filters")
+selected_region = st.sidebar.multiselect(
+    "Select Region (Country)",
+    options=sales_df['LAND1'].unique(),
+    default=sales_df['LAND1'].unique()
+)
+
+filtered_df = sales_df[sales_df['LAND1'].isin(selected_region)]
+
+# --- Main KPIs ---
+st.subheader("Sales KPIs from Real SAP Data")
+col1, col2, col3, col4 = st.columns(4)
+with col1: create_kpi_card("Total Revenue", filtered_df['NETWR'].sum())
+with col2: create_kpi_card("Active Customers", filtered_df['KUNNR'].nunique(), format_str="{:,.0f}")
+with col3: create_kpi_card("Avg Order Value", filtered_df[filtered_df['NETWR'] > 0]['NETWR'].mean())
+with col4: create_kpi_card("Sales Orders", filtered_df['VBELN'].nunique(), format_str="{:,.0f}")
+
+# --- Analytics Tabs ---
+tab1, tab2, tab3, tab4 = st.tabs(["👥 Top Customers", "🌍 Regional Analysis", "📈 Distribution Channels", "🛍️ Top Products"])
+
+with tab1:
+    st.subheader("Top 10 Customers by Revenue")
+    customer_summary = filtered_df.groupby('NAME1')['NETWR'].sum().nlargest(10).reset_index()
+    fig = px.bar(customer_summary, x='NETWR', y='NAME1', orientation='h', labels={'NETWR': 'Revenue ($)', 'NAME1': 'Customer'}, color='NETWR', color_continuous_scale='Blues')
+    st.plotly_chart(fig.update_layout(yaxis={'categoryorder': 'total ascending'}), use_container_width=True)
+
+with tab2:
+    st.subheader("Revenue by Country")
+    regional_summary = filtered_df.groupby('LAND1')['NETWR'].sum().nlargest(10).reset_index()
+    fig = px.pie(regional_summary, values='NETWR', names='LAND1', title="Top 10 Countries by Revenue")
+    st.plotly_chart(fig, use_container_width=True)
+
+with tab3:
+    st.subheader("Revenue by Distribution Channel")
+    channel_summary = filtered_df.groupby('VTWEG')['NETWR'].sum().reset_index()
+    channel_summary['VTWEG'] = channel_summary['VTWEG'].astype(str)
+    fig = px.bar(channel_summary, x='VTWEG', y='NETWR', title="Total Revenue by Distribution Channel", labels={'NETWR': 'Total Revenue ($)', 'VTWEG': 'Distribution Channel'}, color='NETWR', color_continuous_scale='Plasma')
+    st.plotly_chart(fig, use_container_width=True)
+
+with tab4:
+    st.subheader("Top 10 Products by Revenue")
+    product_summary = filtered_df.groupby('MAKTX')['NETWR'].sum().nlargest(10).reset_index()
+    fig = px.bar(product_summary, x='NETWR', y='MAKTX', orientation='h', labels={'NETWR': 'Revenue ($)', 'MAKTX': 'Product'}, color='NETWR', color_continuous_scale='Greens')
+    st.plotly_chart(fig.update_layout(yaxis={'categoryorder': 'total ascending'}), use_container_width=True)
+
+st.markdown("---")
+st.markdown("<p style='text-align: center;'>Built with Streamlit • 100% Real SAP ERP Data from Kaggle</p>", unsafe_allow_html=True)