Update app.py

app.py

@@ -375,8 +375,6 @@ st.markdown(
 )
 
 df_local = df_filtered.copy()
-
-# Tambah kolom bulan
 df_local['created_month'] = df_local['created_at'].dt.to_period('M')
 
 if 'temuan_kode_distrik' in df_local.columns:
@@ -386,146 +384,130 @@ if 'temuan_kode_distrik' in df_local.columns:
         else 'Other'
     )
 
-    # Otomatis bagi dataset berdasarkan Area_Type
     df_pg = df_local[df_local['Area_Type'] == 'PG'].copy()
     df_um = df_local[df_local['Area_Type'] == 'UM'].copy()
 
-    # --- Fungsi untuk menghitung rasio perusahaan ---
+    # --- Hitung rasio per perusahaan ---
     def calculate_avg_ratio_per_company(df_area):
         if df_area.empty:
             return pd.DataFrame()
-        # Hitung temuan per bulan per perusahaan
-        findings_by_company_month = df_area.groupby(['created_month', 'nama_perusahaan']).size().reset_index(name='findings_count')
-        # Hitung jumlah orang unik per bulan per perusahaan
-        creators_by_company_month = df_area.groupby(['created_month', 'nama_perusahaan'])['creator_nid'].nunique().reset_index(name='unique_creators')
-        # Gabung
-        merged = findings_by_company_month.merge(creators_by_company_month, on=['created_month', 'nama_perusahaan'], how='outer')
+        findings = df_area.groupby(['created_month', 'nama_perusahaan']).size().reset_index(name='findings_count')
+        creators = df_area.groupby(['created_month', 'nama_perusahaan'])['creator_nid'].nunique().reset_index(name='unique_creators')
+        merged = findings.merge(creators, on=['created_month', 'nama_perusahaan'], how='outer')
         merged = merged.fillna({'findings_count': 0, 'unique_creators': 0})
-        # Filter untuk menghindari pembagian dengan nol
         merged = merged[merged['unique_creators'] > 0]
-        # Hitung rasio (ignore NaN)
         merged['ratio'] = merged['findings_count'] / merged['unique_creators']
         merged['ratio'] = merged['ratio'].replace([np.inf, -np.inf], np.nan)
 
-        # Jika tidak ada baris valid setelah filter, kembalikan DataFrame kosong
-        if merged.empty:
-            return pd.DataFrame()
-
-        # Rata-rata bulanan per perusahaan
         avg_ratio = merged.groupby('nama_perusahaan')['ratio'].mean().reset_index(name='avg_monthly_ratio')
+        return avg_ratio.dropna(subset=['avg_monthly_ratio'])
 
-        # Jika hasil akhirnya hanya NaN, kembalikan DataFrame kosong
-        if avg_ratio['avg_monthly_ratio'].isna().all():
-            return pd.DataFrame()
-
-        return avg_ratio
-
-    # Hitung untuk masing-masing area
     avg_ratio_pg = calculate_avg_ratio_per_company(df_pg)
     avg_ratio_um = calculate_avg_ratio_per_company(df_um)
 
-    # Palet biru pastel konsisten (soft, harmonis)
-    pastel_blues = [
-        "#A8DADC",  # light cyan
-        "#E2ECE9",  # pale mint
-        "#CCE4E7",  # soft sky
-        "#B5D9D9",  # muted aqua
-        "#98C8D1",  # gentle teal
-        "#7FB9C1",  # calm blue
-        "#6BA9B3"   # deep pastel teal
-    ]
-    pln_color = "#FFD700"  # Kuning PLN
-
-    def assign_colors(df):
-        colors = []
-        blue_idx = 0
-        for company in df['nama_perusahaan']:
-            if 'PLN' in str(company).upper():
-                colors.append(pln_color)
-            else:
-                colors.append(pastel_blues[blue_idx % len(pastel_blues)])
-                blue_idx += 1
-        return colors
+    # 🔑 PALET & COLOR MAPPING GLOBAL — KONSISTEN ANTAR CHART
+    PLN_COLOR = "#FFD700"
+    PASTEL_BLUES = [
+        "#A8DADC", "#E2ECE9", "#CCE4E7", "#B5D9D9", 
+        "#98C8D1", "#7FB9C1", "#6BA9B3", "#5A9CB5", "#4A8FA7", "#3A8399"
+    ]  # diperpanjang sedikit untuk antisipasi banyak kontraktor
 
-    # Fungsi untuk membuat polar bar chart
+    # 🔥 Dapatkan daftar *semua* perusahaan non-PLN unik (dari seluruh data), urut alfabetis → assign warna deterministik
+    all_companies = pd.concat([avg_ratio_pg, avg_ratio_um])['nama_perusahaan'].dropna().unique()
+    non_pln_companies = sorted([c for c in all_companies if 'PLN' not in str(c).upper()])
+    
+    # Mapping: company → color (PLN khusus, lainnya dari palet biru berurut)
+    COMPANY_COLOR_MAP = {}
+    for i, company in enumerate(non_pln_companies):
+        COMPANY_COLOR_MAP[company] = PASTEL_BLUES[i % len(PASTEL_BLUES)]
+    # Pastikan PLN juga masuk mapping (jika muncul)
+    for company in all_companies:
+        if 'PLN' in str(company).upper():
+            COMPANY_COLOR_MAP[company] = PLN_COLOR
+
+    # 🔥 Fungsi helper: dapatkan warna berdasarkan nama perusahaan (konsisten!)
+    def get_color(company_name):
+        return COMPANY_COLOR_MAP.get(company_name, "#CCCCCC")  # fallback abu-abu
+
+    # 🔥 Fungsi chart — pakai mapping global
     def create_polar_bar_chart(df, area_name):
         if df.empty:
             return None
 
-        # Urutkan berdasarkan rasio untuk tampilan konsisten
-        df = df.sort_values('avg_monthly_ratio', ascending=True)
+        # Urutkan untuk stabilitas visual (misal: ascending ratio)
+        df = df.sort_values('avg_monthly_ratio', ascending=True).reset_index(drop=True)
         companies = df['nama_perusahaan'].tolist()
         ratios = df['avg_monthly_ratio'].tolist()
-        colors = assign_colors(df)
-        
-        # Hitung total temuan untuk menghitung sudut (proporsi)
+        colors = [get_color(comp) for comp in companies]
+
+        # Hitung proporsi temuan (untuk lebar bar)
         total_findings = df_local[df_local['Area_Type'] == area_name].groupby('nama_perusahaan').size()
-        angles = []
-        total_angle = 0
-        for comp in companies:
-            count = total_findings.get(comp, 0)
-            angle = (count / total_findings.sum()) * 360 if total_findings.sum() > 0 else 0
-            angles.append(angle)
-        
-        # Hitung sudut tengah untuk setiap bar
+        angles = [total_findings.get(comp, 0) / total_findings.sum() * 360 if total_findings.sum() > 0 else 0 for comp in companies]
+
+        # Hitung posisi tengah
         mid_angles = []
-        current_angle = 0
+        current = 0
         for a in angles:
-            mid_angles.append(current_angle + a / 2)
-            current_angle += a
-        
+            mid_angles.append(current + a / 2)
+            current += a
+
         fig = go.Figure()
-        
-        # 🔥 Tambahkan satu trace untuk setiap perusahaan agar muncul di legend
-        for i, (comp, ratio, color, angle) in enumerate(zip(companies, ratios, colors, angles)):
-               fig.add_trace(go.Barpolar(
-                r=[ratio],
-                theta=[mid_angles[i]],
-                width=[angle],
-                marker_color=[color],
-                marker_line_color="white",
-                marker_line_width=1.2,
-                opacity=0.9,
-                hovertemplate="<b>%{text}</b><br>Avg Ratio: %{r:.2f}<extra></extra>",  # 🔥 Hapus [0]
-                text=[comp],
-                name=comp,
+
+        # Satu trace saja — lebih clean, legend bisa diatur manual jika perlu
+        fig.add_trace(go.Barpolar(
+            r=ratios,
+            theta=mid_angles,
+            width=angles,
+            marker_color=colors,
+            marker_line_color="white",
+            marker_line_width=1.2,
+            opacity=0.9,
+            hovertemplate="<b>%{text}</b><br>Avg Ratio: %{r:.2f}<extra></extra>",
+            text=companies,
+            showlegend=False  # Kita akan buat legend manual yang rapi
+        ))
+
+        # 🔥 LEGEND MANUAL — hanya tampilkan setiap perusahaan sekali, dengan warna konsisten
+        # Tambahkan satu scatter "dummy" per perusahaan untuk legend
+        for company in sorted(set(companies)):
+            color = get_color(company)
+            fig.add_trace(go.Scatterpolar(
+                r=[None], theta=[None],  # invisible
+                mode='markers',
+                marker=dict(color=color, size=10),
+                name=company,
                 showlegend=True
             ))
 
         fig.update_layout(
             title=f'{area_name} Area',
             polar=dict(
-                bgcolor="white",
                 radialaxis=dict(
                     visible=True,
-                    tickfont=dict(size=9, color="gray"),  # 🔥 Warna angka radial jadi putih
+                    tickfont=dict(size=9, color="gray"),
                     gridcolor='lightgray',
-                    title=dict(text='Avg Finding/Person', font=dict(size=10, color="white"))  # 🔥 Warna judul radial
+                    title=dict(text='Avg Finding/Person', font=dict(size=10, color="gray"))
                 ),
                 angularaxis=dict(
                     visible=True,
                     direction='clockwise',
-                    tickfont=dict(size=9, color="white"),  # 🔥 Warna angka derajat jadi putih
+                    tickfont=dict(size=9, color="gray"),
                     showline=False,
-                    linecolor="lightgray",  # 🔥 Warna garis derajat jadi putih
-                    gridcolor="rgba(255,255,255,0.2)"  # 🔥 Warna grid derajat jadi transparan putih
+                    gridcolor="lightgray"
                 ),
             ),
-            showlegend=True,  # 🔥 Aktifkan legend
+            showlegend=True,
             legend=dict(
                 orientation="v",
-                yanchor="top",     # 🔥 Atas
+                yanchor="top",
                 y=1,
-                xanchor="right",   # 🔥 Kanan
+                xanchor="right",
                 x=1.02,
                 font=dict(size=10)
             ),
-            margin=dict(t=40, b=20, l=20, r=20),
-            height=400,
-            paper_bgcolor="rgba(0,0,0,0)",
-            plot_bgcolor="rgba(0,0,0,0)"
+            height=450,
+            margin=dict(t=40, b=20, l=20, r=40)
         )
-
         return fig
 
     # Plot
@@ -536,54 +518,43 @@ if 'temuan_kode_distrik' in df_local.columns:
         fig_pg = create_polar_bar_chart(avg_ratio_pg, 'PG')
         if fig_pg:
             st.plotly_chart(fig_pg, use_container_width=True)
-            
-            # AI Insight untuk PG
             if not avg_ratio_pg.empty:
-                top_company_pg = avg_ratio_pg.loc[avg_ratio_pg['avg_monthly_ratio'].idxmax()]
-                low_company_pg = avg_ratio_pg.loc[avg_ratio_pg['avg_monthly_ratio'].idxmin()]
-                
+                top = avg_ratio_pg.loc[avg_ratio_pg['avg_monthly_ratio'].idxmax()]
+                low = avg_ratio_pg.loc[avg_ratio_pg['avg_monthly_ratio'].idxmin()]
                 st.markdown("### Insight")
-                insight_text = (
+                st.markdown(
                     f"<div class='ai-insight'>"
-                    f"In PG Area, <strong>{top_company_pg['nama_perusahaan']}</strong> has the highest average finding-to-person ratio "
-                    f"(<strong>{top_company_pg['avg_monthly_ratio']:.2f}</strong>), indicating potentially high exposure or active reporting. "
-                    f"Consider reviewing their operational procedures. "
-                    f"Conversely, <strong>{low_company_pg['nama_perusahaan']}</strong> has the lowest ratio "
-                    f"(<strong>{low_company_pg['avg_monthly_ratio']:.2f}</strong>), suggesting the need to actively improve reporting frequency."
-                    f"</div>"
+                    f"In PG Area, <strong>{top['nama_perusahaan']}</strong> has the highest ratio (<strong>{top['avg_monthly_ratio']:.2f}</strong>), "
+                    f"while <strong>{low['nama_perusahaan']}</strong> has the lowest (<strong>{low['avg_monthly_ratio']:.2f}</strong>). "
+                    f"Consider cross-learning between them to standardize reporting culture."
+                    f"</div>",
+                    unsafe_allow_html=True
                 )
-                st.markdown(insight_text, unsafe_allow_html=True)
         else:
-            st.warning("No data for PG area or all ratios are NaN.")
+            st.warning("No data for PG area.")
 
     with col2:
         st.markdown("<h5>Unit Maintenance: Avg Monthly Finding by Company</h5>", unsafe_allow_html=True)
         fig_um = create_polar_bar_chart(avg_ratio_um, 'UM')
         if fig_um:
             st.plotly_chart(fig_um, use_container_width=True)
-            
-            # AI Insight untuk UM
             if not avg_ratio_um.empty:
-                top_company_um = avg_ratio_um.loc[avg_ratio_um['avg_monthly_ratio'].idxmax()]
-                low_company_um = avg_ratio_um.loc[avg_ratio_um['avg_monthly_ratio'].idxmin()]
-                
+                top = avg_ratio_um.loc[avg_ratio_um['avg_monthly_ratio'].idxmax()]
+                low = avg_ratio_um.loc[avg_ratio_um['avg_monthly_ratio'].idxmin()]
                 st.markdown("### Insight")
-                insight_text = (
+                st.markdown(
                     f"<div class='ai-insight'>"
-                    f"In UM Area, all companies show almost the same average finding per person ratio"
-                    # f"(<strong>{top_company_um['avg_monthly_ratio']:.2f}</strong>), warranting a focused safety audit. "
-                    # f"<strong>{low_company_um['nama_perusahaan']}</strong> shows the lowest ratio "
-                    # f"(<strong>{low_company_um['avg_monthly_ratio']:.2f}</strong>), which could reflect strong safety practices or requires verification of reporting completeness."
-                    f"</div>"
+                    f"In UM Area, <strong>{top['nama_perusahaan']}</strong> leads with ratio <strong>{top['avg_monthly_ratio']:.2f}</strong>, "
+                    f"and <strong>{low['nama_perusahaan']}</strong> reports least (<strong>{low['avg_monthly_ratio']:.2f}</strong>). "
+                    f"Check if low reporters have sufficient access to reporting tools or training."
+                    f"</div>",
+                    unsafe_allow_html=True
                 )
-                st.markdown(insight_text, unsafe_allow_html=True)
         else:
-            st.warning("No data for UM area or all ratios are NaN.")
+            st.warning("No data for UM area.")
 else:
-    st.error("Column 'temuan_kode_distrik' not found in the data. Cannot determine PG/UM areas.")
+    st.error("Column 'temuan_kode_distrik' not found.")
     st.stop()
-
-
 # =================== OBJECTIVE 2 — Active vs Inactive Locations (Treemap with Color Gradient) ===================
 st.markdown(
     """