Update app.py

app.py

@@ -41,13 +41,9 @@ def load_data():
 # Load data
 df = load_data()
 
-# Identify price and demand columns
-price_col = [col for col in df.columns if 'price' in col.lower() and col != 'suggested_price']
-demand_col = [col for col in df.columns if 'demand' in col.lower()]
-target_col = 'suggested_price' if 'suggested_price' in df.columns else (price_col[0] if price_col else None)
-
-# Sidebar for navigation
+# Sidebar for navigation and target column selection
 page = st.sidebar.selectbox("Choose a page", ["Data Explorer", "Model Training", "Price Prediction"])
+target_col = st.sidebar.selectbox("Select target column", df.columns.tolist())
 
 if page == "Data Explorer":
     st.title("Data Explorer")
@@ -59,82 +55,81 @@ if page == "Data Explorer":
     st.subheader("Data Visualization")
     fig, ax = plt.subplots(1, 2, figsize=(15, 5))
     
-    if price_col and target_col:
-        ax[0].scatter(df[price_col[0]], df[target_col])
-        ax[0].set_xlabel(price_col[0])
-        ax[0].set_ylabel(target_col)
-        ax[0].set_title(f'{price_col[0]} vs {target_col}')
+    numeric_cols = df.select_dtypes(include=['int64', 'float64']).columns
+    x_col1 = st.selectbox("Select X-axis for first plot", numeric_cols, index=0)
+    x_col2 = st.selectbox("Select X-axis for second plot", numeric_cols, index=min(1, len(numeric_cols)-1))
+    
+    ax[0].scatter(df[x_col1], df[target_col])
+    ax[0].set_xlabel(x_col1)
+    ax[0].set_ylabel(target_col)
+    ax[0].set_title(f'{x_col1} vs {target_col}')
     
-    if demand_col and target_col:
-        ax[1].scatter(df[demand_col[0]], df[target_col])
-        ax[1].set_xlabel(demand_col[0])
-        ax[1].set_ylabel(target_col)
-        ax[1].set_title(f'{demand_col[0]} vs {target_col}')
+    ax[1].scatter(df[x_col2], df[target_col])
+    ax[1].set_xlabel(x_col2)
+    ax[1].set_ylabel(target_col)
+    ax[1].set_title(f'{x_col2} vs {target_col}')
     
     st.pyplot(fig)
 
 elif page == "Model Training":
     st.title("Model Training")
     
-    if not target_col:
-        st.error("No suitable target column (suggested_price or another price column) found in the dataset.")
-    else:
-        # Data preprocessing
-        categorical_cols = df.select_dtypes(include=['object']).columns
-        numeric_cols = df.select_dtypes(include=['int64', 'float64']).columns
-        
-        # Remove target column from features if it exists
-        feature_cols = [col for col in numeric_cols if col != target_col]
-        
-        encoders = {}
-        for col in categorical_cols:
-            encoders[col] = LabelEncoder()
-            df[f'{col}_encoded'] = encoders[col].fit_transform(df[col])
-            feature_cols.append(f'{col}_encoded')
-        
-        X = df[feature_cols]
-        y = df[target_col]
-        
-        scaler = StandardScaler()
-        X_scaled = scaler.fit_transform(X)
+    # Data preprocessing
+    categorical_cols = df.select_dtypes(include=['object']).columns
+    numeric_cols = df.select_dtypes(include=['int64', 'float64']).columns
+    
+    # Remove target column from features
+    feature_cols = [col for col in numeric_cols if col != target_col]
+    
+    encoders = {}
+    for col in categorical_cols:
+        encoders[col] = LabelEncoder()
+        df[f'{col}_encoded'] = encoders[col].fit_transform(df[col])
+        feature_cols.append(f'{col}_encoded')
+    
+    X = df[feature_cols]
+    y = df[target_col]
+    
+    scaler = StandardScaler()
+    X_scaled = scaler.fit_transform(X)
+    
+    # Split the data
+    X_train, X_test, y_train, y_test = train_test_split(X_scaled, y, test_size=0.2, random_state=42)
+    
+    # Model architecture
+    model = Sequential([
+        Dense(64, activation='relu', input_shape=(X_train.shape[1],)),
+        Dense(32, activation='relu'),
+        Dense(16, activation='relu'),
+        Dense(1)
+    ])
+    
+    model.compile(optimizer='adam', loss='mean_squared_error', metrics=['mae'])
+    
+    # Training
+    if st.button("Train Model"):
+        with st.spinner("Training in progress..."):
+            history = model.fit(X_train, y_train, validation_split=0.2, epochs=100, batch_size=32, verbose=0)
         
-        # Split the data
-        X_train, X_test, y_train, y_test = train_test_split(X_scaled, y, test_size=0.2, random_state=42)
+        st.success("Model trained successfully!")
         
-        # Model architecture
-        model = Sequential([
-            Dense(64, activation='relu', input_shape=(X_train.shape[1],)),
-            Dense(32, activation='relu'),
-            Dense(16, activation='relu'),
-            Dense(1)
-        ])
+        # Plot training history
+        fig, ax = plt.subplots(figsize=(10, 5))
+        ax.plot(history.history['loss'], label='Training Loss')
+        ax.plot(history.history['val_loss'], label='Validation Loss')
+        ax.set_xlabel('Epoch')
+        ax.set_ylabel('Loss')
+        ax.legend()
+        st.pyplot(fig)
         
-        model.compile(optimizer='adam', loss='mean_squared_error', metrics=['mae'])
+        # Save model and preprocessing objects
+        model.save('dynamic_pricing_model.h5')
+        joblib.dump(scaler, 'scaler.pkl')
+        joblib.dump(encoders, 'encoders.pkl')
+        joblib.dump(feature_cols, 'feature_cols.pkl')
+        joblib.dump(target_col, 'target_col.pkl')
         
-        # Training
-        if st.button("Train Model"):
-            with st.spinner("Training in progress..."):
-                history = model.fit(X_train, y_train, validation_split=0.2, epochs=100, batch_size=32, verbose=0)
-            
-            st.success("Model trained successfully!")
-            
-            # Plot training history
-            fig, ax = plt.subplots(figsize=(10, 5))
-            ax.plot(history.history['loss'], label='Training Loss')
-            ax.plot(history.history['val_loss'], label='Validation Loss')
-            ax.set_xlabel('Epoch')
-            ax.set_ylabel('Loss')
-            ax.legend()
-            st.pyplot(fig)
-            
-            # Save model and preprocessing objects
-            model.save('dynamic_pricing_model.h5')
-            joblib.dump(scaler, 'scaler.pkl')
-            joblib.dump(encoders, 'encoders.pkl')
-            joblib.dump(feature_cols, 'feature_cols.pkl')
-            joblib.dump(target_col, 'target_col.pkl')
-            
-            st.info("Model and preprocessing objects saved.")
+        st.info("Model and preprocessing objects saved.")
 
 elif page == "Price Prediction":
     st.title("Price Prediction")
@@ -145,7 +140,10 @@ elif page == "Price Prediction":
         scaler = joblib.load('scaler.pkl')
         encoders = joblib.load('encoders.pkl')
         feature_cols = joblib.load('feature_cols.pkl')
-        target_col = joblib.load('target_col.pkl')
+        saved_target_col = joblib.load('target_col.pkl')
+        
+        if saved_target_col != target_col:
+            st.warning(f"Warning: The current target column ({target_col}) is different from the one used for training ({saved_target_col}). Consider retraining the model.")
         
         # User input
         input_data = {}
@@ -157,13 +155,11 @@ elif page == "Price Prediction":
                     input_data[col] = encoders[original_col].transform([value])[0]
             else:
                 if 'year' in col.lower():
-                    input_data[col] = st.slider(f"{col}", 2018, 2024, 2022)
-                elif 'price' in col.lower():
-                    input_data[col] = st.slider(f"{col}", 20.0, 60.0, 40.0)
-                elif 'sales' in col.lower():
-                    input_data[col] = st.slider(f"{col}", 100, 1000, 500)
+                    input_data[col] = st.slider(f"{col}", int(df[col].min()), int(df[col].max()), int(df[col].mean()))
+                elif 'price' in col.lower() or 'sales' in col.lower():
+                    input_data[col] = st.slider(f"{col}", float(df[col].min()), float(df[col].max()), float(df[col].mean()))
                 else:
-                    input_data[col] = st.slider(f"{col}", 0.1, 1.0, 0.5)
+                    input_data[col] = st.slider(f"{col}", float(df[col].min()), float(df[col].max()), float(df[col].mean()))
         
         # Prepare input for prediction
         input_df = pd.DataFrame([input_data])
@@ -172,7 +168,7 @@ elif page == "Price Prediction":
         # Make prediction
         if st.button("Predict Price"):
             predicted_price = model.predict(input_scaled)[0][0]
-            st.success(f"Predicted {target_col}: ${predicted_price:.2f}")
+            st.success(f"Predicted {saved_target_col}: {predicted_price:.2f}")
     else:
         st.warning("Please train the model first!")