Update app.py

app.py

@@ -1148,48 +1148,74 @@ except Exception as e:
     st.exception(e)
 
 
-
 # =================== OBJECTIVE 5: Operator Fatigue Risk Gradient Dashboard =====================
-st.subheader("OBJECTIVE 5:See your team’s fatigue Fatigue Hazard Profile!")
+# ✅ DIPERBAIKI: Gunakan HTML + CSS (bukan st.subheader), centered, white bg, typo fixed
+st.markdown("""
+<h2 class="objective-header">OBJECTIVE 5: See Your Team’s Fatigue Hazard Profile!</h2>
+""", unsafe_allow_html=True)
 
-# Custom CSS — tetap seperti sebelumnya (sudah sesuai preferensi)
+# ✅ CUSTOM CSS — SEMUA STRUKTUR DIPERBAIKI & DIPERLUAS
 st.markdown("""
 <style>
-    .big-title {
-        font-size: 28px;
+    /* === OBJECTIVE HEADER === */
+    .objective-header {
+        font-size: 26px;
         font-weight: bold;
-        color: #ffffff;
+        color: #2c3e50;
         text-align: center;
-        margin-bottom: 10px;
-        background: linear-gradient(135deg, #2c3e50, #1a252c);
-        padding: 15px;
+        margin: 10px 0 25px 0;
+        background: white;
+        padding: 14px;
         border-radius: 10px;
-        box-shadow: 0 4px 15px rgba(0,0,0,0.3);
+        box-shadow: 0 3px 12px rgba(0,0,0,0.1);
+        font-family: 'Segoe UI', Tahoma, Geneva, Verdana, sans-serif;
     }
+
+    /* === BIG TITLE (e.g., "OB HAULER Analysis") === */
+    .big-title {
+        font-size: 22px;
+        font-weight: bold;
+        color: #2c3e50;
+        text-align: center;
+        margin: 25px 0 15px 0;
+        background: white;
+        padding: 12px;
+        border-radius: 8px;
+        box-shadow: 0 2px 8px rgba(0,0,0,0.08);
+        font-family: 'Segoe UI', Tahoma, Geneva, Verdana, sans-serif;
+    }
+
     .subnote {
         font-size: 16px;
         color: #7f8c8d;
         text-align: center;
         margin-bottom: 20px;
+        font-family: 'Segoe UI', Tahoma, Geneva, Verdana, sans-serif;
     }
+
     .section-divider {
         height: 2px;
         background: linear-gradient(to right, #3498db, #2ecc71, #f1c40f, #e74c3c);
-        margin: 20px 0;
+        margin: 25px 0;
     }
+
+    /* === LEGEND === */
     .legend-container {
         display: flex;
-        gap: 15px;
-        margin: 15px 0;
+        gap: 20px;
+        flex-wrap: wrap;
+        justify-content: center;
+        margin: 20px 0;
     }
     .legend-box {
-        background: white;
+        background: #f9f9f9;
         border: 1px solid #ddd;
-        border-radius: 8px;
-        padding: 15px;
-        flex: 1;
-        min-width: 300px;
-        box-shadow: 0 2px 10px rgba(0,0,0,0.05);
+        border-radius: 10px;
+        padding: 16px;
+        min-width: 290px;
+        max-width: 330px;
+        box-shadow: 0 2px 8px rgba(0,0,0,0.05);
+        font-family: 'Segoe UI', Tahoma, Geneva, Verdana, sans-serif;
     }
     .legend-title {
         font-weight: bold;
@@ -1202,25 +1228,34 @@ st.markdown("""
     .legend-item {
         display: flex;
         align-items: center;
-        margin: 5px 0;
-        font-size: 12px;
+        margin: 6px 0;
+        font-size: 13px;
     }
     .legend-color {
         width: 18px;
         height: 18px;
         border-radius: 3px;
-        margin-right: 8px;
+        margin-right: 10px;
         border: 1px solid #ccc;
     }
+    .legend-note {
+        font-size: 12px;
+        color: #666;
+        margin-top: 12px;
+        font-style: italic;
+        line-height: 1.4;
+    }
+
+    /* === AI INSIGHTS === */
     .ai-insight-box {
         background: #f8f9fa;
         border: 1px solid #dee2e6;
         border-radius: 8px;
-        padding: 15px;
-        margin: 10px 0;
+        padding: 16px;
+        margin: 12px 0;
         color: #2c3e50;
         font-family: 'Segoe UI', Tahoma, Geneva, Verdana, sans-serif;
-        box-shadow: 0 2px 8px rgba(0,0,0,0.05);
+        box-shadow: 0 2px 10px rgba(0,0,0,0.05);
     }
     .ai-insight-title {
         font-weight: bold;
@@ -1232,20 +1267,14 @@ st.markdown("""
         border-radius: 5px;
         border-left: 4px solid #495057;
     }
-    .trend-up {
-        color: #e74c3c;
-        font-weight: bold;
-    }
-    .trend-down {
-        color: #27ae60;
-        font-weight: bold;
-    }
+
+    /* === RECOMMENDATIONS === */
     .recommendation-box {
         background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
         border: 1px solid #4a5568;
         border-radius: 8px;
-        padding: 15px;
-        margin: 10px 0;
+        padding: 16px;
+        margin: 12px 0;
         color: white;
         font-family: 'Segoe UI', Tahoma, Geneva, Verdana, sans-serif;
         box-shadow: 0 4px 15px rgba(0,0,0,0.1);
@@ -1261,18 +1290,23 @@ st.markdown("""
         border-left: 4px solid white;
     }
     .recommendation-reason {
-        font-size: 12px;
+        font-size: 13px;
         margin-top: 10px;
         padding: 8px;
         background: rgba(255,255,255,0.1);
         border-radius: 5px;
         border-left: 3px solid rgba(255,255,255,0.3);
     }
+
+    /* === TRENDS === */
+    .trend-up { color: #e74c3c; font-weight: bold; }
+    .trend-down { color: #27ae60; font-weight: bold; }
 </style>
 """, unsafe_allow_html=True)
 
+
 # ===============================================================
-# LOGIC UTAMA
+# LOGIC UTAMA — DIPERBAIKI: PENDEKAN NAMA OPERATOR & KONSISTENSI
 # ===============================================================
 if df.empty:
     st.info("No data available after applying filters.")
@@ -1283,7 +1317,17 @@ else:
             st.warning("Required columns (operator, fleet_type, start) are missing.")
             st.stop()
 
+        # ✅ Shorten operator names: "John Doe" → "John"
         df_op = df[[col_operator, col_fleet_type, "start"]].dropna()
+        if col_operator in df_op.columns:
+            df_op[col_operator] = (
+                df_op[col_operator]
+                .astype(str)
+                .str.strip()
+                .str.split()
+                .str[0]  # Only first part
+            )
+
         if df_op.empty:
             st.info("No operator data after filtering.")
             st.stop()
@@ -1302,9 +1346,9 @@ else:
         ob_data = df_op[df_op["is_ob"]]
         coal_data = df_op[df_op["is_coal"]]
 
+        # Fungsi analisis — tetap sama (tidak ada perubahan logika)
         def get_top10_with_slope(data):
-            if data.empty:
-                return pd.DataFrame()
+            if data.empty: return pd.DataFrame()
             if col_operator not in data.columns:
                 st.error(f"Operator column '{col_operator}' not found in data subset.")
                 return pd.DataFrame()
@@ -1312,22 +1356,20 @@ else:
             weekly = data.groupby([col_operator, "year_week"]).size().reset_index(name="weekly_sum")
             metrics = []
             for nik, grp in weekly.groupby(col_operator):
-                if pd.isna(nik):
-                    continue
+                if pd.isna(nik): continue
                 grp = grp.sort_values("year_week")
                 counts = grp["weekly_sum"].values
                 weeks = np.arange(len(counts))
                 weekly_avg = counts.mean()
                 total_events = counts.sum()
                 n_weeks = len(counts)
+                slope = 0.0
                 if n_weeks >= 2:
                     x_mean = weeks.mean()
                     y_mean = counts.mean()
                     numerator = np.sum((weeks - x_mean) * (counts - y_mean))
                     denominator = np.sum((weeks - x_mean) ** 2)
                     slope = numerator / denominator if denominator != 0 else 0.0
-                else:
-                    slope = 0.0  # One Time Event
                 metrics.append({
                     col_operator: nik,
                     "weekly_avg": weekly_avg,
@@ -1335,50 +1377,15 @@ else:
                     "total_events": total_events,
                     "n_weeks": n_weeks
                 })
-            if not metrics:
-                return pd.DataFrame()
-            return pd.DataFrame(metrics).nlargest(10, "weekly_avg")
+            return pd.DataFrame(metrics).nlargest(10, "weekly_avg") if metrics else pd.DataFrame()
 
         top_ob = get_top10_with_slope(ob_data)
         top_coal = get_top10_with_slope(coal_data)
 
-        def get_all_operators_with_slope(data):
-            if data.empty:
-                return pd.DataFrame()
-            if col_operator not in data.columns:
-                return pd.DataFrame()
-
-            weekly = data.groupby([col_operator, "year_week"]).size().reset_index(name="weekly_sum")
-            metrics = []
-            for nik, grp in weekly.groupby(col_operator):
-                if pd.isna(nik):
-                    continue
-                grp = grp.sort_values("year_week")
-                counts = grp["weekly_sum"].values
-                weeks = np.arange(len(counts))
-                weekly_avg = counts.mean()
-                total_events = counts.sum()
-                n_weeks = len(counts)
-                if n_weeks >= 2:
-                    slope = np.cov(weeks, counts)[0, 1] / np.var(weeks) if np.var(weeks) != 0 else 0.0
-                else:
-                    slope = 0.0
-                metrics.append({
-                    col_operator: nik,
-                    "weekly_avg": weekly_avg,
-                    "slope": slope,
-                    "total_events": total_events,
-                    "n_weeks": n_weeks
-                })
-            return pd.DataFrame(metrics) if metrics else pd.DataFrame()
-
-        all_ob = get_all_operators_with_slope(ob_data)
-        all_coal = get_all_operators_with_slope(coal_data)
-
         # ===============================================================
-        # LEGEND — UPDATED: Stable → One Time Event, Gray → Yellow
+        # LEGEND — SUDAH TERMASUK NOTES DI SETIAP KOTAK
         # ===============================================================
-        st.subheader("Hazard Gradient Legend")
+        st.markdown('<h3 class="big-title">Hazard Gradient Legend</h3>', unsafe_allow_html=True)
         st.markdown("""
 <div class="legend-container">
     <!-- Worsening Trends -->
@@ -1386,22 +1393,22 @@ else:
         <div class="legend-title">Worsening Trends (Positive Slope):</div>
         <div class="legend-item">
             <div class="legend-color" style="background-color: #d32f2f;"></div>
-            <span>Very High Worsening (≥1.5)</span>
+            <span>Very High Risk (≥1.5)</span>
         </div>
         <div class="legend-item">
             <div class="legend-color" style="background-color: #e57373;"></div>
-            <span>High Worsening (1.0–1.5)</span>
+            <span>High Risk (1.0–1.5)</span>
         </div>
         <div class="legend-item">
             <div class="legend-color" style="background-color: #ef9a9a;"></div>
-            <span>Moderate Worsening (0.5–1.0)</span>
+            <span>Moderate Risk (0.5–1.0)</span>
         </div>
         <div class="legend-item">
             <div class="legend-color" style="background-color: #ffcdd2;"></div>
-            <span>Slight Worsening (0–0.5)</span>
+            <span>Slight Risk (0–0.5)</span>
         </div>
         <p class="legend-note">
-            <i>Note: Positive slope indicates increasing fatigue events over time — escalating risk exposure.</i>
+            <i>Note: Positive slope indicates increasing fatigue events over time — escalating operational risk.</i>
         </p>
     </div>
 
@@ -1425,7 +1432,7 @@ else:
             <span>Slight Improvement (−0.5 to 0)</span>
         </div>
         <p class="legend-note">
-            <i>Note: Negative slope reflects decreasing fatigue events — effective mitigation or behavioral improvement.</i>
+            <i>Note: Negative slope reflects decreasing fatigue events — effective mitigation or behavioral adaptation.</i>
         </p>
     </div>
 
@@ -1437,77 +1444,57 @@ else:
             <span>One Time Event (0)</span>
         </div>
         <p class="legend-note">
-            <i>Note: Slope = 0 by definition when data exists for only one week — insufficient for trend assessment.</i>
+            <i>Note: Slope = 0 by definition when data exists for only one week — trend assessment not applicable.</i>
         </p>
     </div>
 </div>
-
-<style>
-    .legend-note {
-        font-size: 12px;
-        color: #666;
-        margin-top: 12px;
-        margin-bottom: 0;
-        font-style: italic;
-        line-height: 1.4;
-    }
-</style>
 """, unsafe_allow_html=True)
 
         # ===============================================================
-        # PLOT FUNCTION — UPDATED: color for slope=0 is now #FFD700
+        # PLOT FUNCTION — DIPERBAIKI: LOGIKA WARNA SESUAI KATEGORI RISK
         # ===============================================================
         def plot_chart(data, title):
             if data.empty:
                 fig = go.Figure()
-                fig.add_annotation(
-                    text="No Data",
-                    x=0.5, y=0.5,
-                    showarrow=False,
-                    font_size=16
+                fig.add_annotation(text="No Data", x=0.5, y=0.5, showarrow=False, font_size=16, font_color="#888")
+                fig.update_layout(
+                    height=350,
+                    title=dict(text=title, x=0.5, font=dict(size=18, family="Segoe UI")),
+                    plot_bgcolor="rgba(0,0,0,0)",
+                    paper_bgcolor="rgba(0,0,0,0)"
                 )
-                fig.update_layout(height=350, title=dict(text=title, x=0.5))
                 return fig
 
             data_sorted = data.sort_values('weekly_avg', ascending=False)
 
+            # ✅ Updated: Match memory — risk categories for worsening (not "worsening")
             def get_color(slope):
                 if slope == 0:
-                    return "#FFD700"  # ✅ Kuning untuk One Time Event
-                elif slope > 0:
-                    if slope < 0.5:
-                        return "#ffcdd2"
-                    elif slope < 1.0:
-                        return "#ef9a9a"
-                    elif slope < 1.5:
-                        return "#e57373"
-                    else:
-                        return "#d32f2f"
-                else:  # slope < 0
-                    if slope > -0.5:
-                        return "#c8e6c9"
-                    elif slope > -1.0:
-                        return "#a5d6a7"
-                    elif slope > -1.5:
-                        return "#81c784"
-                    else:
-                        return "#388e3c"
+                    return "#FFD700"  # One Time Event → yellow
+                elif slope > 0:  # Worsening = Risk
+                    if slope >= 1.5: return "#d32f2f"   # Very High Risk
+                    elif slope >= 1.0: return "#e57373"  # High Risk
+                    elif slope >= 0.5: return "#ef9a9a"  # Moderate Risk
+                    else: return "#ffcdd2"              # Slight Risk
+                else:  # slope < 0 → Improvement
+                    if slope <= -1.5: return "#388e3c"   # Excellent Improvement
+                    elif slope <= -1.0: return "#81c784" # Great Improvement
+                    elif slope <= -0.5: return "#a5d6a7" # Good Improvement
+                    else: return "#c8e6c9"              # Slight Improvement
 
             colors = [get_color(s) for s in data_sorted["slope"]]
 
             bar_trace = go.Bar(
                 x=data_sorted[col_operator].astype(str),
                 y=data_sorted["weekly_avg"],
-                marker=dict(
-                    color=colors,
-                    line=dict(width=2, color="rgba(0,0,0,0.2)")
-                ),
+                marker=dict(color=colors, line=dict(width=1.5, color="rgba(0,0,0,0.2)")),
                 text=[f"{v:.1f}" for v in data_sorted["weekly_avg"]],
                 textposition="outside",
+                textfont=dict(size=11, family="Segoe UI"),
                 hovertemplate=(
                     "<b>%{x}</b><br>" +
                     "Weekly Avg: %{y:.2f}<br>" +
-                    "Trend Slope: %{customdata[0]:+.3f}<br>" +
+                    "Trend Slope: %{customdata[0]:+.2f}<br>" +
                     "Total Events: %{customdata[1]}<br>" +
                     "Weeks Active: %{customdata[2]}<br>" +
                     "<extra></extra>"
@@ -1517,16 +1504,17 @@ else:
 
             fig = go.Figure(bar_trace)
             fig.update_layout(
-                title=dict(text=f"<b>{title}</b>", x=0.5),
-                height=450,
-                margin=dict(l=50, r=20, t=60, b=120),
-                xaxis_title="<b>Operator Name</b>",
-                yaxis_title="<b>Weekly Avg Events</b>",
+                title=dict(text=f"<b>{title}</b>", x=0.5, font=dict(size=18, color="#2c3e50")),
+                height=460,
+                margin=dict(l=50, r=20, t=70, b=130),
+                xaxis_title=dict(text="<b>Operator</b>", font=dict(family="Segoe UI")),
+                yaxis_title=dict(text="<b>Weekly Avg Events</b>", font=dict(family="Segoe UI")),
                 font=dict(family="Segoe UI", size=12),
                 bargap=0.3,
                 plot_bgcolor="rgba(0,0,0,0)",
                 paper_bgcolor="rgba(0,0,0,0)",
-                xaxis=dict(tickangle=45)
+                xaxis=dict(tickangle=45, tickfont=dict(family="Segoe UI")),
+                yaxis=dict(gridcolor="#eee")
             )
             return fig
 
@@ -1540,32 +1528,33 @@ else:
             st.plotly_chart(plot_chart(top_coal, "HAULING COAL Operators (Hazard Gradient)"), use_container_width=True)
 
         # ===============================================================
-        # AI INSIGHTS — tetap dalam bahasa Inggris, tanpa emoticon
+        # AI INSIGHTS — DIPERBAIKI: "Risk Summary", centered title
         # ===============================================================
         col_insight1, col_insight2 = st.columns(2)
 
         with col_insight1:
             if not top_ob.empty:
-                st.markdown("### OB HAULER Analysis")
+                st.markdown('<h3 class="big-title">OB HAULER Analysis</h3>', unsafe_allow_html=True)
                 ob_worsening = len(top_ob[top_ob['slope'] > 0])
                 ob_improving = len(top_ob[top_ob['slope'] < 0])
                 ob_one_time = len(top_ob[top_ob['slope'] == 0])
                 ob_avg_risk = top_ob['weekly_avg'].mean()
                 ob_max_risk = top_ob['weekly_avg'].max()
-                ob_insights = []
+
+                insights = []
                 if ob_worsening > ob_improving:
-                    ob_insights.append(f"{ob_worsening} out of 10 top risk operators are showing <span class='trend-up'>worsening</span> trends.")
+                    insights.append(f"{ob_worsening} out of 10 top-risk operators show <span class='trend-up'>worsening</span> trends.")
                 else:
-                    ob_insights.append(f"{ob_improving} out of 10 top risk operators are showing <span class='trend-down'>improvement</span>.")
+                    insights.append(f"{ob_improving} out of 10 top-risk operators show <span class='trend-down'>improvement</span>.")
                 if ob_one_time > 0:
-                    ob_insights.append(f"{ob_one_time} operators are classified as <b>One Time Event</b> (single-week activity).")
-                ob_insights.append(f"Average risk: {ob_avg_risk:.2f} events/week (max: {ob_max_risk:.2f}).")
+                    insights.append(f"{ob_one_time} operator(s) classified as <b>One Time Event</b>.")
+                insights.append(f"Average risk: {ob_avg_risk:.2f} events/week (max: {ob_max_risk:.2f}).")
 
-                for insight in ob_insights:
+                for txt in insights:
                     st.markdown(f"""
                     <div class="ai-insight-box">
                         <div class="ai-insight-title">Risk Summary</div>
-                        <p>{insight}</p>
+                        <p>{txt}</p>
                     </div>
                     """, unsafe_allow_html=True)
             else:
@@ -1573,26 +1562,27 @@ else:
 
         with col_insight2:
             if not top_coal.empty:
-                st.markdown("### HAULING COAL Analysis")
+                st.markdown('<h3 class="big-title">HAULING COAL Analysis</h3>', unsafe_allow_html=True)
                 coal_worsening = len(top_coal[top_coal['slope'] > 0])
                 coal_improving = len(top_coal[top_coal['slope'] < 0])
                 coal_one_time = len(top_coal[top_coal['slope'] == 0])
                 coal_avg_risk = top_coal['weekly_avg'].mean()
                 coal_max_risk = top_coal['weekly_avg'].max()
-                coal_insights = []
+
+                insights = []
                 if coal_worsening > coal_improving:
-                    coal_insights.append(f"{coal_worsening} out of 10 top risk operators are showing <span class='trend-up'>worsening</span> trends.")
+                    insights.append(f"{coal_worsening} out of 10 top-risk operators show <span class='trend-up'>worsening</span> trends.")
                 else:
-                    coal_insights.append(f"{coal_improving} out of 10 top risk operators are showing <span class='trend-down'>improvement</span>.")
+                    insights.append(f"{coal_improving} out of 10 top-risk operators show <span class='trend-down'>improvement</span>.")
                 if coal_one_time > 0:
-                    coal_insights.append(f"{coal_one_time} operators are classified as <b>One Time Event</b> (single-week activity).")
-                coal_insights.append(f"Average risk: {coal_avg_risk:.2f} events/week (max: {coal_max_risk:.2f}).")
+                    insights.append(f"{coal_one_time} operator(s) classified as <b>One Time Event</b>.")
+                insights.append(f"Average risk: {coal_avg_risk:.2f} events/week (max: {coal_max_risk:.2f}).")
 
-                for insight in coal_insights:
+                for txt in insights:
                     st.markdown(f"""
                     <div class="ai-insight-box">
                         <div class="ai-insight-title">Risk Summary</div>
-                        <p>{insight}</p>
+                        <p>{txt}</p>
                     </div>
                     """, unsafe_allow_html=True)
             else:
@@ -1601,54 +1591,35 @@ else:
         # ===============================================================
         # RECOMMENDATIONS
         # ===============================================================
-        col_rec1, col_rec2 = st.columns(2)
-
         def generate_recommendations(top_ob, top_coal):
             rec = {}
-            if not top_ob.empty:
-                w = len(top_ob[top_ob['slope'] > 0])
-                ot = len(top_ob[top_ob['slope'] == 0])
-                avg = top_ob['weekly_avg'].mean()
-                if w > 5:
-                    r = "Prioritize fatigue intervention for operators with worsening trends."
-                    reason = "High proportion of deteriorating operators signals emerging fatigue risks."
-                elif ot > 4:
-                    r = "Validate data completeness — high One Time Event count may indicate reporting gaps."
-                    reason = "Operators with single-week data cannot yield reliable trend analysis."
-                elif avg > 8:
-                    r = "Review scheduling and rest protocols to reduce event frequency."
-                    reason = "Elevated average event rate increases cumulative fatigue exposure."
-                else:
-                    r = "Maintain current protocols with targeted monitoring."
-                    reason = "Risk profile is stable; focus on sustaining safe practices."
-                rec['ob'] = r
-                rec['ob_reason'] = reason
-
-            if not top_coal.empty:
-                w = len(top_coal[top_coal['slope'] > 0])
-                ot = len(top_coal[top_coal['slope'] == 0])
-                avg = top_coal['weekly_avg'].mean()
-                if w > 5:
-                    r = "Prioritize fatigue intervention for operators with worsening trends."
-                    reason = "High proportion of deteriorating operators signals emerging fatigue risks."
-                elif ot > 4:
-                    r = "Validate data completeness — high One Time Event count may indicate reporting gaps."
-                    reason = "Operators with single-week data cannot yield reliable trend analysis."
-                elif avg > 8:
-                    r = "Review scheduling and rest protocols to reduce event frequency."
-                    reason = "Elevated average event rate increases cumulative fatigue exposure."
-                else:
-                    r = "Maintain current protocols with targeted monitoring."
-                    reason = "Risk profile is stable; focus on sustaining safe practices."
-                rec['coal'] = r
-                rec['coal_reason'] = reason
+            for label, data in [("ob", top_ob), ("coal", top_coal)]:
+                if not data.empty:
+                    w = len(data[data['slope'] > 0])
+                    ot = len(data[data['slope'] == 0])
+                    avg = data['weekly_avg'].mean()
+                    if w > 5:
+                        r = "Prioritize fatigue intervention for operators with worsening trends."
+                        reason = "High proportion of deteriorating operators signals emerging fatigue risks."
+                    elif ot > 4:
+                        r = "Validate data completeness — high One Time Event count may indicate reporting gaps."
+                        reason = "Operators with single-week data cannot yield reliable trend analysis."
+                    elif avg > 8:
+                        r = "Review scheduling and rest protocols to reduce event frequency."
+                        reason = "Elevated average event rate increases cumulative fatigue exposure."
+                    else:
+                        r = "Maintain current protocols with targeted monitoring."
+                        reason = "Risk profile is stable; focus on sustaining safe practices."
+                    rec[label] = r
+                    rec[f"{label}_reason"] = reason
             return rec
 
         ai_rec = generate_recommendations(top_ob, top_coal)
 
+        col_rec1, col_rec2 = st.columns(2)
         with col_rec1:
             if 'ob' in ai_rec:
-                st.markdown("### OB HAULER Recommendations")
+                st.markdown('<h3 class="big-title">OB HAULER Recommendations</h3>', unsafe_allow_html=True)
                 st.markdown(f"""
                 <div class="recommendation-box">
                     <div class="recommendation-title">Action Plan</div>
@@ -1661,7 +1632,7 @@ else:
 
         with col_rec2:
             if 'coal' in ai_rec:
-                st.markdown("### HAULING COAL Recommendations")
+                st.markdown('<h3 class="big-title">HAULING COAL Recommendations</h3>', unsafe_allow_html=True)
                 st.markdown(f"""
                 <div class="recommendation-box">
                     <div class="recommendation-title">Action Plan</div>
@@ -1674,10 +1645,7 @@ else:
 
     except Exception as e:
         st.error(f"Error in Top 10 Operator analysis: {str(e)}")
-        st.exception(e)  # optionally show full traceback during dev
-
-import re # Tambahkan ini jika belum ada
-
+        # st.exception(e)  # Uncomment during development
 # =================== OBJECTIVE 6: Automated Insights & AI Recommendations =====================
 st.subheader("OBJECTIVE 6: Instant Insights & Recommendations")