feat: geographic proximity scoring for website discovery

Prioritize results from Wejherowo region (Norda Biznes home area):
- Wejherowo: +3 pts
- Powiat wejherowski (Reda, Rumia, Luzino...): +2 pts
- Województwo pomorskie: +1 pt
- Outside region: 0 pts

Dashboard shows colored geo badge per candidate.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

blueprints/admin/routes_data_quality.py

@@ -255,6 +255,9 @@ def admin_data_quality():
                 'match_domain': WebsiteDiscoveryService()._domain_matches_company(
                     dc.candidate_domain or '', comp.name
                 ),
+                'match_geo': WebsiteDiscoveryService()._compute_geo_proximity(
+                    dc.page_text_snippet or ''
+                ),
             })
 
         # Count companies without website

services/website_discovery_service.py

@@ -45,6 +45,19 @@ DIRECTORY_DOMAINS = {
     'wikipedia.org', 'olx.pl', 'allegro.pl',
 }
 
+# Geographic proximity scoring for Norda Biznes (Wejherowo region)
+POWIAT_WEJHEROWSKI = {
+    'wejherowo', 'reda', 'rumia', 'luzino', 'gniewino', 'szemud',
+    'łęczyce', 'linia', 'choczewo', 'góra', 'bolszewo', 'gościcino',
+    'nowy dwór wejherowski', 'kąpino', 'bieszkowice', 'sopieszyno',
+}
+WOJEWODZTWO_POMORSKIE = {
+    'gdańsk', 'gdynia', 'sopot', 'słupsk', 'tczew', 'starogard gdański',
+    'chojnice', 'malbork', 'kwidzyn', 'lębork', 'bytów', 'kartuzy',
+    'kościerzyna', 'puck', 'żukowo', 'pruszcz gdański', 'ustka',
+    'władysławowo', 'hel', 'jastarnia', 'łeba',
+}
+
 
 # --- Extraction helpers ---
 
@@ -503,8 +516,35 @@ class WebsiteDiscoveryService:
         else:
             signals['owner'] = False
 
+        # Geographic proximity (weight varies: 3/2/1)
+        signals['geo'] = self._compute_geo_proximity(text, url=None)
+
         return signals
 
+    def _compute_geo_proximity(self, page_text, url=None):
+        """Score geographic proximity to Wejherowo region.
+        Returns: 'wejherowo' (3pt), 'powiat' (2pt), 'pomorskie' (1pt), or False.
+        """
+        text = (page_text or '').lower()
+        if not text:
+            return False
+
+        # Check Wejherowo first (highest priority)
+        if 'wejherowo' in text:
+            return 'wejherowo'
+
+        # Check powiat wejherowski cities
+        for city in POWIAT_WEJHEROWSKI:
+            if city in text:
+                return 'powiat'
+
+        # Check województwo pomorskie
+        for city in WOJEWODZTWO_POMORSKIE:
+            if city in text:
+                return 'pomorskie'
+
+        return False
+
     def _domain_matches_company(self, domain, company_name):
         """Check if domain name matches company name (handles word reordering)."""
         if not domain or not company_name:
@@ -549,6 +589,15 @@ class WebsiteDiscoveryService:
         }
         score = sum(weights.get(k, 0) for k, v in signals.items() if v)
 
+        # Geographic proximity bonus
+        geo = signals.get('geo')
+        if geo == 'wejherowo':
+            score += 3
+        elif geo == 'powiat':
+            score += 2
+        elif geo == 'pomorskie':
+            score += 1
+
         if score >= 5:
             return 'high', score
         elif score >= 2:

templates/admin/data_quality_dashboard.html

@@ -714,6 +714,15 @@
                         <span class="disc-badge {% if d.match_owner %}disc-match{% else %}disc-miss{% endif %}">Właściciel</span>
                         {% endif %}
                         <span class="disc-badge {% if d.match_domain %}disc-match{% else %}disc-miss{% endif %}">Domena</span>
+                        {% if d.match_geo == 'wejherowo' %}
+                        <span class="disc-badge disc-match">Wejherowo</span>
+                        {% elif d.match_geo == 'powiat' %}
+                        <span class="disc-badge disc-match" style="background: #fef3c7; color: #92400e;">Powiat</span>
+                        {% elif d.match_geo == 'pomorskie' %}
+                        <span class="disc-badge disc-match" style="background: #e0e7ff; color: #3730a3;">Pomorskie</span>
+                        {% else %}
+                        <span class="disc-badge disc-miss">Lokalizacja</span>
+                        {% endif %}
                     </div>
                 </td>
                 <td>