Update app.py

app.py

@@ -5,68 +5,97 @@ from sklearn.cluster import KMeans
 from sklearn.preprocessing import StandardScaler
 import matplotlib.pyplot as plt
 import seaborn as sns
+import requests
+from io import StringIO
 
 # App title
 st.title("🛍️ Customer Segmentation Tool")
 
-# Streamlit Tabs
+# 🎯 Streamlit Tabs
 tab1, tab2, tab3, tab4 = st.tabs(["📖 About", "📊 Dataset Overview", "🧑‍🤝‍🧑 Customer Segmentation", "📥 Download Dataset"])
 
 # About Tab
 with tab1:
     st.write("""
-    This app segments customers based on their purchasing behavior using unsupervised learning.
-    You can upload any dataset file type (CSV, Excel) for analysis.
+    This app uses unsupervised learning techniques to segment customers based on their purchasing behavior. 
+    The dataset is uploaded by the user, containing online retail data.
+    ### How It Works:
+    - **Step 1**: Upload customer transaction data, including details like Quantity, UnitPrice, and CustomerID.
+    - **Step 2**: Process the data by calculating the total spent and aggregating the information by customer.
+    - **Step 3**: Apply **K-Means Clustering** to segment the customers into distinct groups.
+    - **Step 4**: Visualize the customer segments with a scatter plot, and optionally download the segmented data.
     """)
 
 # File uploader in the Dataset Tab
 with tab2:
     uploaded_file = st.file_uploader("Upload Your Dataset", type=["csv", "xlsx"])
 
-    def load_data(uploaded_file):
-        if uploaded_file is not None:
-            try:
-                if uploaded_file.name.endswith('.csv'):
-                    df = pd.read_csv(uploaded_file, encoding='ISO-8859-1', on_bad_lines='skip')
-                elif uploaded_file.name.endswith('.xlsx'):
-                    df = pd.read_excel(uploaded_file)
-                return df
-            except Exception as e:
-                st.error(f"Error loading dataset: {e}")
-        return None
+    if uploaded_file is not None:
+        try:
+            # Check file type
+            if uploaded_file.name.endswith('.csv'):
+                # Read CSV file with error handling for malformed lines
+                df = pd.read_csv(uploaded_file, encoding='ISO-8859-1', on_bad_lines='skip')
+            elif uploaded_file.name.endswith('.xlsx'):
+                # Read Excel file
+                df = pd.read_excel(uploaded_file)
+            else:
+                st.error("Unsupported file format. Please upload a CSV or Excel file.")
+                st.stop()
 
-    df = load_data(uploaded_file)
+            st.write("### Dataset Overview")
+            st.write(df.head())
+        except Exception as e:
+            st.error(f"Error loading dataset: {e}")
+            st.stop()
 
-    if df is not None:
-        st.write("### Dataset Overview")
-        st.write(df.head())
+        # Automatically detect possible columns
+        st.write("### Columns detected in your dataset:")
+        st.write(df.columns.tolist())
 
-# Customer Segmentation Tab
-with tab3:
-    if df is not None:
-        # Column selection
-        st.write("### Select Columns")
-        customer_col = st.selectbox("Select Customer Column", df.columns.tolist(), index=0)
-        quantity_col = st.selectbox("Select Quantity Column", df.columns.tolist(), index=0)
-        unit_price_col = st.selectbox("Select Unit Price Column", df.columns.tolist(), index=0)
-        
-        df = df.dropna(subset=[customer_col])
-        df["TotalSpent"] = pd.to_numeric(df[quantity_col], errors='coerce') * pd.to_numeric(df[unit_price_col], errors='coerce')
+        # Allow the user to map columns
+        customer_col = st.selectbox("Select Customer Column", df.columns.tolist(), index=df.columns.tolist().index("CustomerID") if "CustomerID" in df.columns else 0)
+        quantity_col = st.selectbox("Select Quantity Column", df.columns.tolist(), index=df.columns.tolist().index("Quantity") if "Quantity" in df.columns else 0)
+        unit_price_col = st.selectbox("Select Unit Price Column", df.columns.tolist(), index=df.columns.tolist().index("UnitPrice") if "UnitPrice" in df.columns else 0)
+
+        # Check if the selected columns exist
+        if customer_col not in df.columns or quantity_col not in df.columns or unit_price_col not in df.columns:
+            st.error("One or more selected columns do not exist in the dataset. Please select valid columns.")
+            st.stop()
+
+        # Preprocess data
+        df = df.dropna(subset=[customer_col])  # Remove rows without CustomerID
+        df["TotalSpent"] = pd.to_numeric(df[quantity_col], errors='coerce') * pd.to_numeric(df[unit_price_col], errors='coerce')  # Ensure numeric type
+
+        # Ensure that TotalSpent column is not NaN
         df = df.dropna(subset=["TotalSpent"])
 
+        # Aggregate data by Customer
         customer_data = df.groupby(customer_col).agg({
             "TotalSpent": "sum",
             quantity_col: "sum",
             unit_price_col: "mean"
         }).rename(columns={quantity_col: "NumTransactions", unit_price_col: "AvgUnitPrice"})
 
+        # Debug: Check if 'NumTransactions' exists in the DataFrame
+        st.write("### Available columns in the aggregated customer data:")
+        st.write(customer_data.columns.tolist())
+
+        # Standardize the data
         scaler = StandardScaler()
         customer_scaled = pd.DataFrame(scaler.fit_transform(customer_data), columns=customer_data.columns, index=customer_data.index)
 
+# Customer Segmentation Tab
+with tab3:
+    if uploaded_file is not None:
+        # User selects the number of clusters
         num_clusters = st.slider("Select Number of Clusters", min_value=2, max_value=10, value=3)
+
+        # Apply K-Means clustering
         model = KMeans(n_clusters=num_clusters, random_state=42)
         customer_data["Cluster"] = model.fit_predict(customer_scaled)
 
+        # Visualize the clusters
         st.write("### Clusters Visualization")
         fig, ax = plt.subplots()
         scatter = ax.scatter(customer_data["TotalSpent"], customer_data["NumTransactions"], c=customer_data["Cluster"], cmap='viridis')
@@ -76,7 +105,52 @@ with tab3:
         plt.colorbar(scatter, label="Cluster")
         st.pyplot(fig)
 
+        # Show the segmented customer data
+        st.write("### Customer Segments Data")
+        st.write(customer_data.head())
+
+        # Option to download the segmented data
         csv = customer_data.to_csv(index=True)
-        st.download_button("Download Segmented Customer Data", data=csv, file_name="segmented_customer_data.csv", mime="text/csv")
+        st.download_button(
+            label="Download Segmented Customer Data",
+            data=csv,
+            file_name="segmented_customer_data.csv",
+            mime="text/csv"
+        )
     else:
         st.write("Please upload a dataset to start.")
+
+# Download Dataset Tab
+with tab4:
+    st.write("""
+    You can download the sample 'Online Retail' dataset to get started with customer segmentation tasks.
+    Click the button below to download the dataset in CSV format.
+    """)
+
+    # Direct Google Drive link to the 'Online Retail' dataset (for direct download)
+    dataset_url_online_retail = "https://drive.google.com/uc?id=1djBqO2sdHfy9DGZQXZu2Er8LUUXtp9Kr&export=download"
+
+    # Direct Google Drive link to the new dataset (for direct download)
+    dataset_url_new_file = "https://drive.google.com/uc?id=1PbGJSdcyDInsu-9Ua4iHzQh-YpVk_RqT&export=download"
+
+    # Download the file from the URLs
+    response_online_retail = requests.get(dataset_url_online_retail)
+    file_data_online_retail = response_online_retail.text  # Get the content as text
+
+    response_new_file = requests.get(dataset_url_new_file)
+    file_data_new_file = response_new_file.text  # Get the content as text
+
+    # Convert the CSV data into a CSV download for Streamlit
+    st.download_button(
+        label="Download Online Retail Dataset",
+        data=file_data_online_retail,
+        file_name="Online_Retail.csv",
+        mime="text/csv"
+    )
+    
+    st.download_button(
+        label="Download New Dataset",
+        data=file_data_new_file,
+        file_name="New_Dataset.csv",
+        mime="text/csv"
+    )