diff --git a/test_data/data-linestring-encoding_native.parquet b/test_data/data-linestring-encoding_native.parquet
new file mode 100644
index 0000000..264705d
Binary files /dev/null and b/test_data/data-linestring-encoding_native.parquet differ
diff --git a/test_data/data-linestring-encoding_wkb.parquet b/test_data/data-linestring-encoding_wkb.parquet
new file mode 100644
index 0000000..fd8b1d4
Binary files /dev/null and b/test_data/data-linestring-encoding_wkb.parquet differ
diff --git a/test_data/data-linestring-wkt.csv b/test_data/data-linestring-wkt.csv
new file mode 100644
index 0000000..a3a320d
--- /dev/null
+++ b/test_data/data-linestring-wkt.csv
@@ -0,0 +1,4 @@
+"col","geometry"
+0,"LINESTRING (30 10, 10 30, 40 40)"
+1,"LINESTRING EMPTY"
+2,
diff --git a/test_data/data-multilinestring-encoding_native.parquet b/test_data/data-multilinestring-encoding_native.parquet
new file mode 100644
index 0000000..2bb5822
Binary files /dev/null and b/test_data/data-multilinestring-encoding_native.parquet differ
diff --git a/test_data/data-multilinestring-encoding_wkb.parquet b/test_data/data-multilinestring-encoding_wkb.parquet
new file mode 100644
index 0000000..08f9f0d
Binary files /dev/null and b/test_data/data-multilinestring-encoding_wkb.parquet differ
diff --git a/test_data/data-multilinestring-wkt.csv b/test_data/data-multilinestring-wkt.csv
new file mode 100644
index 0000000..1c5f1ea
--- /dev/null
+++ b/test_data/data-multilinestring-wkt.csv
@@ -0,0 +1,5 @@
+"col","geometry"
+0,"MULTILINESTRING ((30 10, 10 30, 40 40))"
+1,"MULTILINESTRING ((10 10, 20 20, 10 40), (40 40, 30 30, 40 20, 30 10))"
+2,"MULTILINESTRING EMPTY"
+3,
diff --git a/test_data/data-multipoint-encoding_native.parquet b/test_data/data-multipoint-encoding_native.parquet
new file mode 100644
index 0000000..b0435ef
Binary files /dev/null and b/test_data/data-multipoint-encoding_native.parquet differ
diff --git a/test_data/data-multipoint-encoding_wkb.parquet b/test_data/data-multipoint-encoding_wkb.parquet
new file mode 100644
index 0000000..8454f25
Binary files /dev/null and b/test_data/data-multipoint-encoding_wkb.parquet differ
diff --git a/test_data/data-multipoint-wkt.csv b/test_data/data-multipoint-wkt.csv
new file mode 100644
index 0000000..00d926f
--- /dev/null
+++ b/test_data/data-multipoint-wkt.csv
@@ -0,0 +1,5 @@
+"col","geometry"
+0,"MULTIPOINT ((30 10))"
+1,"MULTIPOINT ((10 40), (40 30), (20 20), (30 10))"
+2,"MULTIPOINT EMPTY"
+3,
diff --git a/test_data/data-multipolygon-encoding_native.parquet b/test_data/data-multipolygon-encoding_native.parquet
new file mode 100644
index 0000000..43a8d2f
Binary files /dev/null and b/test_data/data-multipolygon-encoding_native.parquet differ
diff --git a/test_data/data-multipolygon-encoding_wkb.parquet b/test_data/data-multipolygon-encoding_wkb.parquet
new file mode 100644
index 0000000..538b8f4
Binary files /dev/null and b/test_data/data-multipolygon-encoding_wkb.parquet differ
diff --git a/test_data/data-multipolygon-wkt.csv b/test_data/data-multipolygon-wkt.csv
new file mode 100644
index 0000000..211a681
--- /dev/null
+++ b/test_data/data-multipolygon-wkt.csv
@@ -0,0 +1,6 @@
+"col","geometry"
+0,"MULTIPOLYGON (((30 10, 40 40, 20 40, 10 20, 30 10)))"
+1,"MULTIPOLYGON (((30 20, 45 40, 10 40, 30 20)), ((15 5, 40 10, 10 20, 5 10, 15 5)))"
+2,"MULTIPOLYGON (((40 40, 20 45, 45 30, 40 40)), ((20 35, 10 30, 10 10, 30 5, 45 20, 20 35), (30 20, 20 15, 20 25, 30 20)))"
+3,"MULTIPOLYGON EMPTY"
+4,
diff --git a/test_data/data-point-encoding_native.parquet b/test_data/data-point-encoding_native.parquet
new file mode 100644
index 0000000..4e6489a
Binary files /dev/null and b/test_data/data-point-encoding_native.parquet differ
diff --git a/test_data/data-point-encoding_wkb.parquet b/test_data/data-point-encoding_wkb.parquet
new file mode 100644
index 0000000..94b87d1
Binary files /dev/null and b/test_data/data-point-encoding_wkb.parquet differ
diff --git a/test_data/data-point-wkt.csv b/test_data/data-point-wkt.csv
new file mode 100644
index 0000000..8164a4b
--- /dev/null
+++ b/test_data/data-point-wkt.csv
@@ -0,0 +1,5 @@
+"col","geometry"
+0,"POINT (30 10)"
+1,"POINT EMPTY"
+2,
+3,"POINT (40 40)"
diff --git a/test_data/data-polygon-encoding_native.parquet b/test_data/data-polygon-encoding_native.parquet
new file mode 100644
index 0000000..68d1aa6
Binary files /dev/null and b/test_data/data-polygon-encoding_native.parquet differ
diff --git a/test_data/data-polygon-encoding_wkb.parquet b/test_data/data-polygon-encoding_wkb.parquet
new file mode 100644
index 0000000..cce77ba
Binary files /dev/null and b/test_data/data-polygon-encoding_wkb.parquet differ
diff --git a/test_data/data-polygon-wkt.csv b/test_data/data-polygon-wkt.csv
new file mode 100644
index 0000000..0574892
--- /dev/null
+++ b/test_data/data-polygon-wkt.csv
@@ -0,0 +1,5 @@
+"col","geometry"
+0,"POLYGON ((30 10, 40 40, 20 40, 10 20, 30 10))"
+1,"POLYGON ((35 10, 45 45, 15 40, 10 20, 35 10), (20 30, 35 35, 30 20, 20 30))"
+2,"POLYGON EMPTY"
+3,
diff --git a/test_data/generate_test_data.py b/test_data/generate_test_data.py
new file mode 100644
index 0000000..0e50a14
--- /dev/null
+++ b/test_data/generate_test_data.py
@@ -0,0 +1,218 @@
+"""
+Generates example data using pyarrow by running `python generate_test_data.py`.
+
+You can print the metadata with:
+
+.. code-block:: python
+
+   >>> import json, pprint, pyarrow.parquet as pq
+   >>> pprint.pprint(json.loads(pq.read_schema("example.parquet").metadata[b"geo"]))
+"""
+import json
+import pathlib
+import copy
+
+import numpy as np
+import pyarrow as pa
+import pyarrow.parquet as pq
+from pyarrow.csv import write_csv
+
+from shapely import from_wkt, to_wkb
+
+
+HERE = pathlib.Path(__file__).parent
+
+
+metadata_template = {
+    "version": "1.1.0",
+    "primary_column": "geometry",
+    "columns": {
+        "geometry": {
+            "encoding": "WKB",
+            "geometry_types": [],
+        },
+    },
+}
+
+
+## Various geometry types with WKB and native (GeoArrow-based) encodings
+
+def write_encoding_files(geometries_wkt, geometries_geoarrow, geometry_type):
+
+    table = pa.table({"col": range(len(geometries_wkt)), "geometry": geometries_wkt})
+    write_csv(table, HERE / f"data-{geometry_type.lower()}-wkt.csv")
+
+    # WKB encoding
+    table = pa.table(
+        {"col": range(len(geometries_wkt)), "geometry": to_wkb(from_wkt(geometries_wkt))}
+    )
+    metadata = copy.deepcopy(metadata_template)
+    metadata["columns"]["geometry"]["geometry_types"] = [geometry_type]
+    table = table.replace_schema_metadata({"geo": json.dumps(metadata)})
+    pq.write_table(table, HERE / f"data-{geometry_type.lower()}-encoding_wkb.parquet")
+
+    # native (geoarrow) encoding
+    table = pa.table(
+        {"col": range(len(geometries_wkt)), "geometry": geometries_geoarrow}
+    )
+    metadata["columns"]["geometry"]["encoding"] = geometry_type.lower()
+    table = table.replace_schema_metadata({"geo": json.dumps(metadata)})
+    pq.write_table(table, HERE / f"data-{geometry_type.lower()}-encoding_native.parquet")
+
+
+# point
+
+geometries_wkt = [
+    "POINT (30 10)",
+    "POINT EMPTY",
+    None,
+    "POINT (40 40)",
+]
+
+point_type = pa.struct(
+    [
+        pa.field("x", pa.float64(), nullable=False),
+        pa.field("y", pa.float64(), nullable=False)
+    ]
+)
+geometries = pa.array(
+    [(30, 10), (float("nan"), float("nan")), (float("nan"), float("nan")), (40, 40)],
+    mask=np.array([False, False, True, False]),
+    type=point_type
+)
+
+write_encoding_files(
+    geometries_wkt, geometries, geometry_type="Point"
+)
+
+# linestring
+
+geometries_wkt = [
+    "LINESTRING (30 10, 10 30, 40 40)",
+    "LINESTRING EMPTY",
+    None
+]
+
+linestring_type = pa.list_(pa.field("vertices", point_type, nullable=False))
+geometries = pa.array(
+    [[(30, 10), (10, 30), (40, 40)], [], []],
+    mask=np.array([False, False, True]),
+    type=linestring_type
+)
+
+write_encoding_files(
+    geometries_wkt, geometries, geometry_type="LineString"
+)
+
+# polygon
+
+geometries_wkt = [
+    "POLYGON ((30 10, 40 40, 20 40, 10 20, 30 10))",
+    "POLYGON ((35 10, 45 45, 15 40, 10 20, 35 10), (20 30, 35 35, 30 20, 20 30))",
+    "POLYGON EMPTY",
+    None,
+]
+
+polygon_type = pa.list_(
+    pa.field("rings", pa.list_(
+        pa.field("vertices", point_type, nullable=False)
+    ), nullable=False)
+)
+geometries = pa.array(
+    [
+        [[(30, 10), (40, 40), (20, 40), (10, 20), (30, 10)]],
+        [[(35, 10), (45, 45), (15, 40), (10, 20), (35, 10)],
+         [(20, 30), (35, 35), (30, 20), (20, 30)]],
+        [],
+        [],
+    ],
+    mask=np.array([False, False, False, True]),
+    type=polygon_type
+)
+
+write_encoding_files(
+    geometries_wkt, geometries, geometry_type="Polygon"
+)
+
+# multipoint
+
+geometries_wkt = [
+    "MULTIPOINT ((30 10))",
+    "MULTIPOINT ((10 40), (40 30), (20 20), (30 10))",
+    "MULTIPOINT EMPTY",
+    None,
+]
+
+multipoint_type = pa.list_(pa.field("points", point_type, nullable=False))
+geometries = pa.array(
+    [
+        [(30, 10)],
+        [(10, 40), (40, 30), (20, 20), (30, 10)],
+        [],
+        [],
+    ],
+    mask=np.array([False, False, False, True]),
+    type=multipoint_type
+)
+
+write_encoding_files(
+    geometries_wkt, geometries, geometry_type="MultiPoint"
+)
+
+# multilinestring
+
+geometries_wkt = [
+    "MULTILINESTRING ((30 10, 10 30, 40 40))",
+    "MULTILINESTRING ((10 10, 20 20, 10 40), (40 40, 30 30, 40 20, 30 10))",
+    "MULTILINESTRING EMPTY",
+    None,
+]
+
+multilinestring_type = pa.list_(
+    pa.field("linestrings", linestring_type, nullable=False)
+)
+geometries = pa.array(
+    [
+        [[(30, 10), (10, 30), (40, 40)]],
+        [[(10, 10), (20, 20), (10, 40)],
+         [(40, 40), (30, 30), (40, 20), (30, 10)]],
+        [],
+        [],
+    ],
+    mask=np.array([False, False, False, True]),
+    type=multilinestring_type
+)
+
+write_encoding_files(
+    geometries_wkt, geometries, geometry_type="MultiLineString"
+)
+
+# multipolygon
+
+geometries_wkt = [
+    "MULTIPOLYGON (((30 10, 40 40, 20 40, 10 20, 30 10)))",
+    "MULTIPOLYGON (((30 20, 45 40, 10 40, 30 20)), ((15 5, 40 10, 10 20, 5 10, 15 5)))",
+    "MULTIPOLYGON (((40 40, 20 45, 45 30, 40 40)), ((20 35, 10 30, 10 10, 30 5, 45 20, 20 35), (30 20, 20 15, 20 25, 30 20)))",
+    "MULTIPOLYGON EMPTY",
+    None,
+]
+
+multipolygon_type = pa.list_(pa.field("polygons", polygon_type, nullable=False))
+geometries = pa.array(
+    [
+        [[[(30, 10), (40, 40), (20, 40), (10, 20), (30, 10)]]],
+        [[[(30, 20), (45, 40), (10, 40), (30, 20)]],
+         [[(15, 5), (40, 10), (10, 20), (5, 10), (15, 5)]]],
+        [[[(40, 40), (20, 45), (45, 30), (40, 40)]],
+         [[(20, 35), (10, 30), (10, 10), (30, 5), (45, 20), (20, 35)],
+          [(30, 20), (20, 15), (20, 25), (30, 20)]]],
+        [],
+        [],
+    ],
+    mask=np.array([False, False, False, False, True]),
+    type=multipolygon_type
+)
+
+write_encoding_files(
+    geometries_wkt, geometries, geometry_type="MultiPolygon"
+)