diff --git a/Dockerfile b/Dockerfile
index 5b188e15..55e67e28 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -6,7 +6,7 @@ RUN apk add nodejs yarn git python3 python3-dev py3-pip \
     py3-cairo file
 
 # Download and install Tippecanoe
-RUN git clone -b 1.35.0 https://github.com/mapbox/tippecanoe.git /tmp/tippecanoe && \
+RUN git clone -b 1.36.0 https://github.com/mapbox/tippecanoe.git /tmp/tippecanoe && \
     cd /tmp/tippecanoe && \
     make && \
     PREFIX=/usr/local make install && \
diff --git a/openaddr/conform.py b/openaddr/conform.py
index 01db7b39..ef65fc10 100644
--- a/openaddr/conform.py
+++ b/openaddr/conform.py
@@ -869,10 +869,15 @@ def _transform_to_4326(srs):
     if srs not in _transform_cache:
         epsg_id = int(srs[5:]) if srs.startswith("EPSG:") else int(srs)
         # Manufacture a transform object if it's not in the cache
+
         in_spatial_ref = osr.SpatialReference()
         in_spatial_ref.ImportFromEPSG(epsg_id)
         out_spatial_ref = osr.SpatialReference()
         out_spatial_ref.ImportFromEPSG(4326)
+
+        # GDAL 3 changes axis order: https://github.com/OSGeo/gdal/issues/1546
+        out_spatial_ref.SetAxisMappingStrategy(osgeo.osr.OAMS_TRADITIONAL_GIS_ORDER)
+
         _transform_cache[srs] = osr.CoordinateTransformation(in_spatial_ref, out_spatial_ref)
     return _transform_cache[srs]
 
@@ -937,6 +942,7 @@ def row_extract_and_reproject(source_config, source_row):
             out_row[GEOM_FIELDNAME] = None
             return out_row
 
+
     # Reproject the coordinates if necessary
     if "srs" in data_source["conform"]:
         try:
diff --git a/openaddr/tests/__init__.py b/openaddr/tests/__init__.py
index 2d657c54..1346cde9 100644
--- a/openaddr/tests/__init__.py
+++ b/openaddr/tests/__init__.py
@@ -901,9 +901,9 @@ def test_single_jp_fukushima2(self):
         self.assertEqual(rows[1]['STREET'], u'田沢字躑躅ケ森')
         self.assertEqual(rows[2]['NUMBER'], u'22-9')
         self.assertEqual(rows[2]['STREET'], u'小田字正夫田')
-        self.assertEqual(rows[0]['GEOM'], 'POINT (140.480007 37.706391)')
-        self.assertEqual(rows[1]['GEOM'], 'POINT (140.486267 37.707664)')
-        self.assertEqual(rows[2]['GEOM'], 'POINT (140.41875 37.710239)')
+        self.assertEqual(rows[0]['GEOM'], 'POINT (37.706391 140.480007)')
+        self.assertEqual(rows[1]['GEOM'], 'POINT (37.707664 140.486267)')
+        self.assertEqual(rows[2]['GEOM'], 'POINT (37.710239 140.41875)')
 
     def test_single_utah(self):
         ''' Test complete process_one.process on data that uses file selection with mixed case (issue #104)
diff --git a/openaddr/tests/conform.py b/openaddr/tests/conform.py
index d53973fe..0940f620 100644
--- a/openaddr/tests/conform.py
+++ b/openaddr/tests/conform.py
@@ -550,7 +550,7 @@ def test_row_extract_and_reproject(self):
         }), "addresses", "default")
         r = row_extract_and_reproject(d, {GEOM_FIELDNAME: "POINT (7655634.924 668868.414)"})
 
-        self.assertEqual('POINT (45.4815543938511 -122.630842186651)', r[GEOM_FIELDNAME])
+        self.assertEqual('POINT (-122.630842186651 45.4815543938511)', r[GEOM_FIELDNAME])
 
         d = SourceConfig(dict({
             "schema": 2,
@@ -1840,7 +1840,7 @@ def test_lake_man_3740(self):
         self.assertEqual(0, rc)
         with open(dest_path) as fp:
             rows = list(csv.DictReader(fp))
-            self.assertEqual(rows[0]['GEOM'], 'POINT (37.8026123 -122.2592495)')
+            self.assertEqual(rows[0]['GEOM'], 'POINT (-122.2592495 37.8026123)')
             self.assertEqual(rows[0]['NUMBER'], '5')
             self.assertEqual(rows[0]['STREET'], u'PZ ESPA\u00d1A')
 
@@ -2212,7 +2212,7 @@ def test_srs(self):
              u'3203,SE WOODSTOCK BLVD,7655634.924,668868.414'.encode('ascii'))
         r = self._convert(c, d)
         self.assertEqual(r[0], u'n,s,{GEOM_FIELDNAME}'.format(**globals()))
-        self.assertEqual(r[1], u'3203,SE WOODSTOCK BLVD,POINT (45.4815543938511 -122.630842186651)')
+        self.assertEqual(r[1], u'3203,SE WOODSTOCK BLVD,POINT (-122.630842186651 45.4815543938511)')
 
     def test_too_many_columns(self):
         "Check that we don't barf on input with too many columns in some rows"
diff --git a/openaddr/tests/conforms/jp-nara.json b/openaddr/tests/conforms/jp-nara.json
index 88c88332..88326755 100644
--- a/openaddr/tests/conforms/jp-nara.json
+++ b/openaddr/tests/conforms/jp-nara.json
@@ -15,7 +15,7 @@
             },
             "conform": {
                 "format": "csv",
-                "lat": "\u7def\u5ea6",
+                "lon": "\u7def\u5ea6",
                 "srs": "EPSG:4612",
                 "number": "auto_number",
                 "number": {
@@ -27,7 +27,7 @@
                     "separator": "-"
                 },
                 "street": "\u5927\u5b57\u30fb\u753a\u4e01\u76ee\u540d",
-                "lon": "\u7d4c\u5ea6",
+                "lat": "\u7d4c\u5ea6",
                 "encoding": "SHIFT_JISX0213"
             },
             "license": "http://nlftp.mlit.go.jp/ksj/other/yakkan.html"
diff --git a/setup.py b/setup.py
index 4fe21125..f4f9ed1b 100644
--- a/setup.py
+++ b/setup.py
@@ -64,7 +64,6 @@
 
         # Used in openaddr.parcels
         'Shapely == 1.7.1',
-        'Fiona == 1.8.17',
 
         # Used in dotmaps preview to support S3-backed SQLite mbtiles
         # https://rogerbinns.github.io/apsw/
@@ -77,6 +76,5 @@
         'mapbox-vector-tile==1.2.0',
         'future==0.16.0',
         'protobuf==3.5.1',
-        'pyclipper==1.1.0'
         ]
 )