Update app.py

app.py

@@ -16,7 +16,6 @@ def build_dataframe_from_inputs(values):
     for name, minutes, variety in values:
         minutes = 0.0 if minutes is None else float(minutes)
         variety = None if variety is None else float(variety)
-        # Keep row if there is any meaningful input
         if minutes > 0 or (variety is not None and not np.isnan(variety)):
             rows.append(
                 {
@@ -63,7 +62,6 @@ def analyze_user_data(
     if df.empty:
         return "Please enter at least one platform with some weekly minutes.", pd.DataFrame()
 
-    # Call core model
     scores = cear_analyzer.calculate_scores(
         df, satisfaction=feed_satisfaction, fomo=fomo_level
     )
@@ -76,11 +74,11 @@ def analyze_user_data(
     fomo = scores.get("FOMO", None)
     per_eff = scores.get("Per_Platform_Efficiency", [])
 
-    # Profile from C and A
+    # --- profile based on C & A --- #
     if c >= 70 and a >= 70:
         profile = (
             "You are highly plugged into online culture, but that comes with high "
-            "algorithmic risk and a heavy concentration of attention."
+            "algorithmic risk and a heavy concentration of attention on a small set of feeds."
         )
     elif c >= 70 and a < 70:
         profile = (
@@ -90,15 +88,15 @@ def analyze_user_data(
     elif c < 40 and a >= 70:
         profile = (
             "You give a lot of attention to a narrow set of feeds without gaining much cultural exposure. "
-            "This is a classic high-risk, low-benefit pattern."
+            "This is a high-risk, low-benefit pattern."
         )
     else:
         profile = (
             "You currently have relatively low exposure to viral trends and also keep algorithmic risk low. "
-            "You are either deliberately detached or under-invested in highly trend-dense platforms."
+            "You are either deliberately detached from viral culture or simply under-invested in trend-dense platforms."
         )
 
-    # Variety interpretation
+    # --- variety interpretation --- #
     if avg_variety is None:
         variety_text = (
             "You did not provide variety ratings, so this analysis focuses only on time and platform mix."
@@ -106,7 +104,7 @@ def analyze_user_data(
     elif avg_variety < 4:
         variety_text = (
             f"Your average variety rating is **{avg_variety:.1f} / 10**, which suggests that your feeds feel "
-            "quite repetitive and may reinforce a narrow slice of content."
+            "quite repetitive and reinforce a narrow slice of content."
         )
     elif avg_variety > 7:
         variety_text = (
@@ -118,7 +116,7 @@ def analyze_user_data(
             f"Your average variety rating is **{avg_variety:.1f} / 10**, indicating a moderate mix of content types."
         )
 
-    # Satisfaction/FOMO interpretation
+    # --- satisfaction & FOMO interpretation --- #
     satisfaction_text = ""
     if satisfaction is not None:
         if satisfaction <= 3:
@@ -128,12 +126,13 @@ def analyze_user_data(
             )
         elif satisfaction >= 8:
             satisfaction_text = (
-                "You report high satisfaction with your feed, indicating your current pattern largely feels aligned "
-                "with your preferences."
+                "You report high satisfaction with your feed, indicating your current usage largely aligns "
+                "with what you want out of these platforms."
             )
         else:
             satisfaction_text = (
-                "Your satisfaction is in the middle range, which suggests room for improvement without a complete overhaul."
+                "Your satisfaction is in the mid range, which suggests your feed is 'fine' but not fully optimized "
+                "for how you would like to spend your attention."
             )
 
     fomo_text = ""
@@ -141,7 +140,7 @@ def analyze_user_data(
         if fomo >= 7 and c < 50:
             fomo_text = (
                 "You feel out of the loop and your relatively low C-Score supports that feeling. "
-                "A bit more time on trend-dense platforms could help if staying current matters to you."
+                "If staying current matters to you, a bit more time on trend-dense platforms could help."
             )
         elif fomo <= 3 and c < 40:
             fomo_text = (
@@ -149,6 +148,7 @@ def analyze_user_data(
                 "distance from viral culture."
             )
 
+    # --- summary header --- #
     summary_lines = [
         "## 📊 CEAR Analysis Summary",
         "",
@@ -157,18 +157,13 @@ def analyze_user_data(
         f"- **Platform Diversity Index (D-Index):** **{d:.2f}**",
     ]
     if avg_variety is not None:
-        summary_lines.append(
-            f"- **Average Variety Rating (0–10):** **{avg_variety:.2f}**"
-        )
+        summary_lines.append(f"- **Average Variety Rating (0–10):** **{avg_variety:.2f}**")
     if satisfaction is not None:
-        summary_lines.append(
-            f"- **Feed Satisfaction (0–10):** **{satisfaction:.1f}**"
-        )
+        summary_lines.append(f"- **Feed Satisfaction (0–10):** **{satisfaction:.1f}**")
     if fomo is not None:
-        summary_lines.append(
-            f"- **FOMO / Out-of-the-loop feeling (0–10):** **{fomo:.1f}**"
-        )
+        summary_lines.append(f"- **FOMO / Out-of-the-loop (0–10):** **{fomo:.1f}**")
 
+    # --- interpretation section --- #
     summary_lines.extend(
         [
             "",
@@ -186,6 +181,18 @@ def analyze_user_data(
         summary_lines.append("")
         summary_lines.append(fomo_text)
 
+    # --- explanation of how survey inputs are used --- #
+    survey_explainer = """
+### ℹ️ How your answers are used
+
+- **Minutes per week** drive the core scores. More time on high-weight platforms increases both C-Score and A-Risk, with diminishing returns for C-Score.
+- **Per-platform variety (0–10)** is combined into a minutes-weighted average. Low variety means you mainly see one type of content; high variety means you see a wider mix of topics and styles.
+- **Feed satisfaction (0–10)** does not change the scores; it is used to interpret whether your current pattern feels good or frustrating to you.
+- **FOMO (0–10)** is compared with your C-Score: high FOMO with low C-Score means you feel out of the loop, while low FOMO with low C-Score means you are detached by choice.
+"""
+    summary_lines.append("")
+    summary_lines.append(survey_explainer.strip())
+
     summary_lines.append(
         "\nThe C-Score uses a logarithmic transform of your weekly minutes, encoding diminishing returns as time increases. "
         "A-Risk reflects your raw time investment and how concentrated it is on a small set of high-weight platforms. "
@@ -194,13 +201,14 @@ def analyze_user_data(
 
     summary = "\n".join(summary_lines).strip()
 
-    # Per-platform efficiency table
+    # --- per-platform efficiency table --- #
     if isinstance(per_eff, list) and per_eff:
         eff_df = pd.DataFrame(per_eff)
         if "platform_name" in eff_df.columns:
             eff_df = eff_df.rename(
                 columns={"platform_name": "platform", "Cultural_Efficiency": "efficiency_score"}
             )
+        eff_df["efficiency_score"] = eff_df["efficiency_score"].round(1)
         eff_df = eff_df.sort_values("efficiency_score", ascending=False)
     else:
         eff_df = pd.DataFrame(columns=["platform", "efficiency_score"])
@@ -208,7 +216,7 @@ def analyze_user_data(
     return summary, eff_df
 
 
-# ---------------- Gradio UI ----------------
+# ---------------- Gradio UI ---------------- #
 
 with gr.Blocks() as demo:
     gr.Markdown(