Merge pull request #342 from Swapnilden/correct_prediction

Fix Model Evaluation Metrics Display Issue

Author: sanjay-kv

Job Satisfaction Analysis/app.py

@@ -1,9 +1,7 @@
 import streamlit as st
 import joblib
 import pandas as pd
-import matplotlib.pyplot as plt
-import seaborn as sns
-from sklearn.metrics import accuracy_score, confusion_matrix, classification_report, roc_curve, auc
+from sklearn.preprocessing import LabelEncoder
 
 # Load the model
 model = joblib.load('model.pkl')
@@ -12,6 +10,26 @@
 features = ['Hobby', 'OpenSource', 'Country', 'Student', 'Employment', 'FormalEducation',
             'UndergradMajor', 'CompanySize', 'DevType', 'YearsCoding', 'YearsCodingProf']
 
+# Initialize label encoders for categorical features
+encoders = {
+    'Hobby': LabelEncoder().fit(['Yes', 'No']),
+    'OpenSource': LabelEncoder().fit(['Yes', 'No']),
+    'Country': LabelEncoder().fit(['United States', 'India', 'Germany']),
+    'Student': LabelEncoder().fit(['Yes', 'No']),
+    'Employment': LabelEncoder().fit(['Employed full-time', 'Employed part-time', 'Self-employed', 'Unemployed']),
+    'FormalEducation': LabelEncoder().fit(["Bachelor’s degree (BA, BS, B.Eng., etc.)", 
+                                           "Master’s degree (MA, MS, M.Eng., MBA, etc.)", 
+                                           "Doctoral degree (PhD)"]),
+    'UndergradMajor': LabelEncoder().fit(["Computer science, computer engineering, or software engineering", 
+                                          "Information technology, networking, or system administration", 
+                                          "Other engineering discipline"]),
+    'CompanySize': LabelEncoder().fit(['Fewer than 10 employees', '10 to 19 employees', '20 to 99 employees', 
+                                       '100 to 499 employees', '500 to 999 employees', '1,000 to 4,999 employees']),
+    'DevType': LabelEncoder().fit(['Developer, back-end', 'Developer, front-end', 'Developer, full-stack']),
+    'YearsCoding': LabelEncoder().fit(['0-2 years', '3-5 years', '6-8 years', '9-11 years']),
+    'YearsCodingProf': LabelEncoder().fit(['0-2 years', '3-5 years', '6-8 years', '9-11 years']),
+}
+
 st.title('Job Satisfaction Prediction')
 
 # Create a form for user input
@@ -54,6 +72,10 @@
     # Convert user input to DataFrame
     input_df = pd.DataFrame([input_data])
 
+    # Encode categorical features
+    for feature in features:
+        input_df[feature] = encoders[feature].transform(input_df[feature])
+    
     # Ensure the input has the same columns as the training data
     input_df = input_df[features]
 
@@ -62,47 +84,3 @@
 
     # Display the prediction
     st.write(f'Predicted Job Satisfaction: {prediction[0]}')
-
-    # Evaluate the model on test data (assuming y_test and y_pred are available)
-    # This part would typically be done during model development, not in the prediction app
-    # However, for demonstration purposes, we can create some dummy data
-    y_test = [1, 0, 1, 1, 0]  # Example true labels
-    y_pred = model.predict(input_df)  # Example predicted labels
-
-    # Print accuracy
-    accuracy = accuracy_score(y_test, y_pred)
-    st.write(f'Accuracy: {accuracy:.2f}')
-
-    # Print classification report
-    report = classification_report(y_test, y_pred, output_dict=True)
-    st.write('Classification Report:')
-    st.write(report)
-
-    # Convert classification report to a DataFrame for better readability
-    report_df = pd.DataFrame(report).transpose()
-    st.write(report_df)
-
-    # Plot confusion matrix
-    cm = confusion_matrix(y_test, y_pred)
-    plt.figure(figsize=(10, 6))
-    sns.heatmap(cm, annot=True, fmt='d', cmap='Blues', cbar=False)
-    plt.title('Confusion Matrix')
-    plt.xlabel('Predicted')
-    plt.ylabel('Actual')
-    st.pyplot(plt)
-
-    # If the model is a binary classifier, plot the ROC curve
-    if len(set(y_test)) == 2:
-        fpr, tpr, _ = roc_curve(y_test, y_pred)
-        roc_auc = auc(fpr, tpr)
-        
-        plt.figure(figsize=(10, 6))
-        plt.plot(fpr, tpr, color='darkorange', lw=2, label=f'ROC curve (area = {roc_auc:.2f})')
-        plt.plot([0, 1], [0, 1], color='navy', lw=2, linestyle='--')
-        plt.xlim([0.0, 1.0])
-        plt.ylim([0.0, 1.05])
-        plt.xlabel('False Positive Rate')
-        plt.ylabel('True Positive Rate')
-        plt.title('Receiver Operating Characteristic (ROC) Curve')
-        plt.legend(loc='lower right')
-        st.pyplot(plt)