Show radar chart for Likert cluters; add tabs

page_ai.py

@@ -5,23 +5,36 @@ from matplotlib.font_manager import FontProperties
 import matplotlib.pyplot as plt
 import seaborn as sns
 import numpy as np
+from sklearn.cluster import KMeans
+from sklearn.decomposition import PCA
 from fields.prod_feat_flat_fields import prod_feat_flat_fields
 from fields.feature_translations import feature_translations
+from fields.likert_flat_fields import likert_flat_fields
 
 #@st.cache_data
 def show(df):
     # Load the Chinese font
     chinese_font = FontProperties(fname='notosans.ttf', size=12)
     st.title("AI Companion")
-    st.write("Clustering students based on AI-assistant feature choices")
-    clusters = perform_kmodes_clustering(df, prod_feat_flat_fields)
-    st.markdown(
-                f"<h2 style='text-align: center;'>Feature Preferences</h2>", unsafe_allow_html=True)
-    show_radar_chart(clusters, font_prop=chinese_font)
-    plot_feature_preferences(clusters, font_prop=chinese_font)
-    st.markdown(
-                f"<h2 style='text-align: center;'>Preferred AI Roles</h2>", unsafe_allow_html=True)
-    visualize_ai_roles(df, chinese_font)
+
+    tab1, tab2 = st.tabs(["Likert-Based Clustering", "Feature-Based Clustering"])
+
+    with tab1:
+        st.write("AI-assistant feature choices per Likert-based Personas")
+        likert_cluster_and_visualize(df, likert_flat_fields, chinese_font)
+
+    with tab2:
+        st.write("Clustering students based on AI-assistant feature choices")
+        clusters = perform_kmodes_clustering(df, prod_feat_flat_fields)
+        st.markdown(
+                    f"<h2 style='text-align: center;'>Feature Preferences (Overall)</h2>", unsafe_allow_html=True)
+        show_radar_chart(clusters, font_prop=chinese_font)
+        st.markdown(
+                    f"<h2 style='text-align: center;'>Feature Preferences (By Cluster)</h2>", unsafe_allow_html=True)
+        plot_feature_preferences(clusters, font_prop=chinese_font)
+        st.markdown(
+                    f"<h2 style='text-align: center;'>Preferred AI Roles (Overall)</h2>", unsafe_allow_html=True)
+        visualize_ai_roles(df, chinese_font)
 
 
 def visualize_ai_roles(df, chinese_font):
@@ -35,7 +48,7 @@ def visualize_ai_roles(df, chinese_font):
     # Plot the data
     plt.figure(figsize=(10, 6))
     ai_roles_data.plot(kind='bar', color='skyblue')
-    plt.title('Desired AI Roles', fontproperties=chinese_font)
+    plt.title('Preferred AI Roles', fontproperties=chinese_font)
     plt.xlabel('Roles', fontproperties=chinese_font)
     plt.ylabel('Number of Responses', fontproperties=chinese_font)
     plt.xticks(rotation=45, ha='right', fontproperties=chinese_font)
@@ -174,3 +187,79 @@ def plot_feature_preferences(clusters, font_prop):
 
     # Streamlit uses st.pyplot() to display matplotlib charts
     st.pyplot(fig)
+
+def likert_cluster_and_visualize(df, likert_flat_fields, chinese_font):
+    # Clean the DataFrame column names
+    df.columns = [col.strip() for col in df.columns]
+
+    # Also clean the likert_flat_fields if necessary
+    likert_flat_fields = [field.strip() for field in likert_flat_fields]
+
+    # Prepare the likert data, dropping any rows with missing values
+    df_likert_data = df[likert_flat_fields].dropna()
+
+    # Perform k-means clustering
+    kmeans = KMeans(n_clusters=3, n_init=10, random_state=42).fit(df_likert_data)
+    df_likert_data['Cluster'] = kmeans.labels_
+
+    # Concatenate the cluster labels with the original data
+    df_clustered = pd.concat([df, df_likert_data['Cluster']], axis=1)
+
+    # Aggregate the product preference data for each cluster
+    cluster_preferences = []
+    for i in range(3):
+        cluster_data = df_clustered[df_clustered['Cluster'] == i]
+        cluster_preferences.append(cluster_data[prod_feat_flat_fields].mean())
+
+    # Radar Chart Plotting
+    df_dict = {
+        'Eco-Friendly': cluster_preferences[0],
+        'Moderate': cluster_preferences[1],
+        'Frugal': cluster_preferences[2]
+    }
+
+    feature_translations_dict = dict(zip(prod_feat_flat_fields, feature_translations))
+    persona_averages = [df_dict[key].tolist() for key in df_dict]
+
+    # Append the first value at the end of each list for the radar chart
+    for averages in persona_averages:
+        averages += averages[:1]
+    
+    # Prepare the English labels for plotting
+    english_feature_labels = list(feature_translations)
+    english_feature_labels += [english_feature_labels[0]]  # Repeat the first label to close the loop
+    
+    # Number of variables we're plotting
+    num_vars = len(english_feature_labels)
+    
+    # Split the circle into even parts and save the angles
+    angles = np.linspace(0, 2 * np.pi, num_vars, endpoint=False).tolist()
+    angles += angles[:1]  # Complete the loop
+    
+    # Set up the font properties for using a custom font
+    fig, ax = plt.subplots(figsize=(12, 12), subplot_kw=dict(polar=True))
+    fig.subplots_adjust(left=0.1, right=0.9, top=0.9, bottom=0.1)
+
+    # Draw one axe per variable and add labels
+    plt.xticks(angles[:-1], english_feature_labels, color='grey', size=12, fontproperties=chinese_font)
+    
+    # Draw ylabels
+    ax.set_rlabel_position(0)
+    plt.yticks([0.2, 0.4, 0.6, 0.8, 1], ["0.2", "0.4", "0.6", "0.8", "1"], color="grey", size=7)
+    plt.ylim(0, 1)
+    
+    # Plot data and fill with color
+    for label, data in zip(df_dict.keys(), persona_averages):
+        data += data[:1]  # Complete the loop
+        ax.plot(angles, data, label=label, linewidth=1, linestyle='solid')
+        ax.fill(angles, data, alpha=0.25)
+    
+    # Add legend
+    plt.legend(title='Personas')
+    plt.legend(loc='upper right', bbox_to_anchor=(0.1, 0.1))
+    
+    # Add a title
+    plt.title('Product Feature Preferences by Persona', size=20, color='grey', y=1.1, fontproperties=chinese_font)
+    
+    # Display the radar chart
+    st.pyplot(fig)