Update app.py

app.py

@@ -2,11 +2,10 @@ import gradio as gr
 from cear_model import CEARModel
 import pandas as pd
 
-# Instantiate the model globally
+# Initialize Model
 cear_analyzer = CEARModel()
 
-# --- Configuration: Default Values for Pre-population ---
-# Users will see these rows filled in when the app loads
+# --- Configuration: Default Values ---
 DEFAULT_DATA = [
     ["TikTok", 0],
     ["Instagram", 0],
@@ -17,85 +16,124 @@ DEFAULT_DATA = [
     ["LinkedIn", 0]
 ]
 
-def analyze_user_data(input_table, variety_rating, fomo_rating):
+def interpret_scores(c_score, a_risk, d_index):
     """
-    Analyzes screen time + subjective survey data.
+    Converts raw numbers into human-readable context.
     """
-    # 1. Convert Gradio input to DataFrame
+    # --- 1. Interpret Algorithmic Risk (A-Risk) ---
+    # Thresholds: <150 (Low), 150-500 (Medium), >500 (High)
+    if a_risk < 150:
+        a_status = "🟢 Low Risk"
+        a_desc = "Your usage is within a safe range, likely minimizing algorithmic manipulation."
+    elif a_risk < 500:
+        a_status = "🟡 Moderate Risk"
+        a_desc = "You have significant exposure to algorithmic feeds. Be mindful of 'rabbit holes'."
+    else:
+        a_status = "🔴 High Risk"
+        a_desc = "Your attention is heavily concentrated in high-risk algorithmic environments."
+
+    # --- 2. Interpret Cultural Connectedness (C-Score) ---
+    # Thresholds: <2.0 (Disconnected), 2.0-5.0 (Connected), >5.0 (Very High)
+    if c_score < 2.0:
+        c_status = "⚪ Low Exposure"
+        c_desc = "You likely miss many viral trends, which protects focus but reduces cultural sync."
+    elif c_score < 5.0:
+        c_status = "🔵 Balanced Exposure"
+        c_desc = "You are plugged into main trends without being overwhelmed."
+    else:
+        c_status = "🟣 High Saturation"
+        c_desc = "You are extremely plugged in. Ensure this doesn't lead to information overload."
+
+    # --- 3. Interpret Diversity (D-Index) ---
+    # Thresholds: 1.0 (Monoculture) to >3.0 (Diverse)
+    if d_index < 1.5:
+        d_status = "⚠️ Concentrated"
+        d_desc = "Your time is spent almost entirely on one platform (Echo Chamber risk)."
+    elif d_index < 3.0:
+        d_status = "⚖️ Moderate Spread"
+        d_desc = "You split time across a few apps, offering some variety."
+    else:
+        d_status = "🌐 High Diversity"
+        d_desc = "Your usage is well-distributed, reducing reliance on any single algorithm."
+
+    return {
+        "A_Status": a_status, "A_Desc": a_desc,
+        "C_Status": c_status, "C_Desc": c_desc,
+        "D_Status": d_status, "D_Desc": d_desc
+    }
+
+def analyze_user_data(input_table, variety_rating, fomo_rating):
+    # 1. Process Input
     user_data_df = pd.DataFrame(input_table, columns=['platform_name', 'minutes_per_week'])
-    
-    # Clean inputs: coerce non-numbers to 0
     user_data_df['minutes_per_week'] = pd.to_numeric(user_data_df['minutes_per_week'], errors='coerce').fillna(0)
 
-    # 2. Call the core model
+    # 2. Run Model
     raw_scores = cear_analyzer.calculate_scores(user_data_df)
-
-    # 3. Generate Analysis Summary
-    # We now integrate the survey scores into the interpretation
     
-    # Logic for simple feedback based on the new survey inputs
-    variety_status = "High" if variety_rating >= 4 else ("Low" if variety_rating <= 2 else "Moderate")
-    fomo_status = "High" if fomo_rating >= 4 else ("Low" if fomo_rating <= 2 else "Moderate")
+    # 3. Get Interpretation
+    context = interpret_scores(raw_scores['C_Score'], raw_scores['A_Risk'], raw_scores['D_Index'])
 
+    # 4. Generate Rich Markdown Output
     summary = f"""
-    ## 📊 Analysis Summary
-    
-    ### 🧠 Core Metrics
-    - **Cultural Connectedness Score (C-Score):** **{raw_scores['C_Score']:.2f}**
-    - **Algorithmic Risk Score (A-Risk):** **{raw_scores['A_Risk']:.2f}**
-    - **Platform Diversity Index (D-Index):** **{raw_scores['D_Index']:.2f}**
+    ## 📊 Analysis Results
 
-    ### 🗣️ Self-Reported Context
-    - **Perceived Variety:** {variety_rating}/5 ({variety_status})
-    - **FOMO Intensity:** {fomo_rating}/5 ({fomo_status})
+    ### 1. Algorithmic Risk: {context['A_Status']} 
+    *Score: {raw_scores['A_Risk']:.0f}*
+    > {context['A_Desc']}
 
-    ---
-    ### 📝 Interpretation
-    *Your C-Score reflects exposure to viral trends based on logarithmically scaled time. Your A-Risk captures vulnerability to high-intensity feeds.*
+    ### 2. Cultural Connectedness: {context['C_Status']}
+    *Score: {raw_scores['C_Score']:.2f}*
+    > {context['C_Desc']}
+
+    ### 3. Platform Diversity: {context['D_Status']}
+    *Index: {raw_scores['D_Index']:.2f}*
+    > {context['D_Desc']}
     
-    **Insight:** You reported **{variety_status} variety** and **{fomo_status} FOMO**. 
-    {'⚠️ High FOMO suggests you may be relying on these feeds for social validation despite the algorithmic risk.' if fomo_status == 'High' else '✅ Low FOMO indicates a healthy detachment from the feed cycle.'}
+    ---
+    ### 🗣️ Self-Reported Context
+    * **Variety:** {variety_rating}/5
+    * **FOMO:** {fomo_rating}/5
     """
 
-    # Return the summary text and the efficiency table
     return summary, pd.DataFrame(raw_scores['Per_Platform_Efficiency'])
 
-# --- Define the Gradio Interface ---
-with gr.Blocks(title="CEAR Model") as demo:
+# --- Interface Definition ---
+with gr.Blocks(title="CEAR Model", theme=gr.themes.Soft()) as demo:
     gr.Markdown("# 🧭 Cultural Exposure & Algorithmic Risk (CEAR) Model")
-    gr.Markdown("Enter your weekly screen time (minutes) and rate your experience below.")
+    gr.Markdown("""
+    **Instructions:**
+    1. Enter your estimated **minutes per week** for each app (e.g., check your phone's Screen Time).
+    2. Adjust the sliders for your personal experience.
+    3. Click **Analyze** to see if your habits align with your goals.
+    """)
 
     with gr.Row():
         with gr.Column(scale=2):
-            # The Dataframe with pre-populated values
             input_table = gr.Dataframe(
                 headers=['platform_name', 'minutes_per_week'],
-                value=DEFAULT_DATA,  # <--- THIS PRE-POPULATES THE ROWS
+                value=DEFAULT_DATA,
                 datatype=["str", "number"],
                 row_count=7,
                 col_count=(2, 'fixed'),
-                label="Weekly Screen Time (Minutes)"
+                label="Weekly Screen Time Inputs"
             )
         
         with gr.Column(scale=1):
-            # New Survey Questions
             gr.Markdown("### 🧠 Self-Reflection")
-            variety_slider = gr.Slider(1, 5, step=1, value=3, label="How diverse does your feed feel? (1=Repetitive, 5=Diverse)")
-            fomo_slider = gr.Slider(1, 5, step=1, value=3, label="How strong is your FOMO? (1=None, 5=Intense)")
+            variety_slider = gr.Slider(1, 5, step=1, value=3, label="Perceived Variety (1=Samey, 5=Diverse)")
+            fomo_slider = gr.Slider(1, 5, step=1, value=3, label="FOMO Intensity (1=None, 5=High)")
             
-            analyze_btn = gr.Button("Analyze My Risk", variant="primary")
+            analyze_btn = gr.Button("🔍 Analyze Risks", variant="primary")
 
     with gr.Row():
-        output_text = gr.Markdown(label="Analysis Results")
+        output_text = gr.Markdown(label="Detailed Report")
         output_table = gr.Dataframe(label="Cultural Efficiency (Score per Minute)")
 
-    # Connect the inputs to the function
     analyze_btn.click(
         fn=analyze_user_data,
         inputs=[input_table, variety_slider, fomo_slider],
         outputs=[output_text, output_table]
     )
 
-    # Launch with the required server binding
     if __name__ == "__main__":
         demo.launch(server_name="0.0.0.0", server_port=7860)