diff --git a/process_osm.py b/process_osm.py
index d0f1d67..bc57c51 100644
--- a/process_osm.py
+++ b/process_osm.py
@@ -8,12 +8,14 @@
 from src.boundaries import BoundariesWriter
 from src.buildings import BuildingsWriter
 from src.settlements import SettlementsWriter
+from src.parks import ParksWriter
 
 LAYERS = {
     "highways": HighwaysWriter,
     "boundaries": BoundariesWriter,
     "buildings": BuildingsWriter,
     "settlements": SettlementsWriter,
+    "parks": ParksWriter,
 }
 
 
diff --git a/src/parks.py b/src/parks.py
new file mode 100644
index 0000000..42de6fe
--- /dev/null
+++ b/src/parks.py
@@ -0,0 +1,84 @@
+import sys
+
+import pyarrow
+from osmium.osm import TagList
+
+from .geoparquet import GeoParquetWriter
+from .helpers import tags_with_prefix, split_multi_value_field
+
+
+class ParksWriter(GeoParquetWriter):
+    COLUMNS = [
+        ("boundary", pyarrow.string()),
+        ("protected_area", pyarrow.string()),
+        ("leisure", pyarrow.string()),
+        ("name", pyarrow.list_(pyarrow.string())),
+        ("names", pyarrow.map_(pyarrow.string(), pyarrow.list_(pyarrow.string()))),
+        ("short_name", pyarrow.list_(pyarrow.string())),
+        ("short_names", pyarrow.map_(pyarrow.string(), pyarrow.list_(pyarrow.string()))),
+        ("official_name", pyarrow.list_(pyarrow.string())),
+        ("official_names", pyarrow.map_(pyarrow.string(), pyarrow.list_(pyarrow.string()))),
+        ("protect_class", pyarrow.string()),
+        ("protection_title", pyarrow.string()),
+        ("protected", pyarrow.string()),
+        ("iucn_level", pyarrow.string()),
+        ("access", pyarrow.string()),
+        ("operator", pyarrow.string()),
+        ("operator:type", pyarrow.string()),
+        ("owner", pyarrow.string()),
+        ("ownership", pyarrow.string()),
+        ("start_date", pyarrow.string()),
+        ("related_law", pyarrow.string()),
+        ("website", pyarrow.string()),
+        ("wikidata", pyarrow.string()),
+        ("wikipedia", pyarrow.string()),
+    ]
+
+    FILTERS = {"boundary", "leisure"}
+
+    def node(self, o):
+        if not self._is_park(o.tags):
+            return
+
+        try:
+            self.append(
+                "node",
+                o.id,
+                self.columns(o.tags),
+                self.wkbfactory.create_point(o),
+            )
+        except RuntimeError as e:
+            print(e, file=sys.stderr)
+
+    def area(self, o):
+        if not self._is_park(o.tags):
+            return
+
+        try:
+            self.append(
+                "way" if o.from_way() else "relation",
+                o.orig_id(),
+                self.columns(o.tags),
+                self.wkbfactory.create_multipolygon(o),
+            )
+        except RuntimeError as e:
+            print(e, file=sys.stderr)
+
+    def _is_park(self, tags):
+        return (
+            tags.get("boundary") in {"national_park", "protected_area"} or
+            tags.get("leisure") in {"park", "nature_reserve"}
+        )
+
+    def columns(self, tags: TagList):
+        return {key: column(key, tags) for (key, _) in self.COLUMNS}
+
+
+def column(column_name: str, tags: TagList):
+    match column_name:
+        case "name" | "short_name" | "official_name":
+            return split_multi_value_field(tags.get(column_name))
+        case "names" | "short_names" | "official_names":
+            return tags_with_prefix(f"{column_name[:-1]}:", tags, transform=split_multi_value_field)
+        case _:
+            return tags.get(column_name)