diff --git a/python/housing_data/build_metros.py b/python/housing_data/build_metros.py
index 7e11137..733aa50 100644
--- a/python/housing_data/build_metros.py
+++ b/python/housing_data/build_metros.py
@@ -5,7 +5,7 @@
 
 
 def load_crosswalk_df(data_repo_path: Path) -> pd.DataFrame:
-    crosswalk_df = pd.read_csv(data_repo_path / "data/crosswalk/cbsa2fipsxw.csv")
+    crosswalk_df = pd.read_csv(data_repo_path / "data/crosswalk/cbsa2fipsxw_2023.csv")
 
     # Drop the μSAs, no one cares about them.
     # Most of them are just one county anyway, so showing the combined metro stats doesn't
diff --git a/python/housing_data/california_hcd_data.py b/python/housing_data/california_hcd_data.py
index 8601ca7..8e1c429 100644
--- a/python/housing_data/california_hcd_data.py
+++ b/python/housing_data/california_hcd_data.py
@@ -76,7 +76,9 @@ def load_california_hcd_data(
         ],
         None,
     )
-    assert df["building_type"].isnull().sum() == 0
+
+    assert df["building_type"].isnull().sum() < 50
+    df = df[df["building_type"].notnull()]
 
     df = df.rename(columns={"YEAR": "year"}).astype({"year": str})