Fix Washington, DC places and counties view

python/housing_data/build_data_utils.py

@@ -183,7 +183,15 @@ def get_state_abbrs(state_codes: pd.Series) -> pd.Series:
     :param state_codes: state_codes: pd.Series of int
     :return: pd.Series of state abbrs as str
     """
-    return state_codes.astype(str).str.zfill(2).map(us.states.mapping("fips", "abbr"))
+    # (see https://en.wikipedia.org/wiki/Federal_Information_Processing_Standard_state_code#FIPS_state_codes).
+    fips_to_abbr = us.states.mapping("fips", "abbr") | {
+        # DC is not in us.states.STATES_AND_TERRITORIES for some reason
+        "11": "DC",
+        # BPS uses alternate FIPS codes for Puerto Rico and Virgin Islands, idk why
+        "43": "PR",
+        "52": "VI",
+    }
+    return state_codes.astype(str).str.zfill(2).map(fips_to_abbr)
 
 
 def load_bps_all_years_plus_monthly(

python/housing_data/build_places.py

@@ -325,7 +325,10 @@ def load_places(
     places_df["path_2"] = name.str.replace("/", "-").str.replace(" ", "_")
     places_df = places_df.drop(columns=["place_name", "place_type"])
 
-    # Not sure why I have to do this
-    places_df = places_df[places_df["path_1"].notnull() & places_df["path_2"].notnull()]
+    not_null_rows = places_df["path_1"].notnull() & places_df["path_2"].notnull()
+    assert (
+        not_null_rows.all()
+    ), f"Found rows where path_1 or path_2 is null: {places_df[['path_1', 'path_2']]}"
+    places_df = places_df[not_null_rows]
 
     return raw_places_df, places_df

python/housing_data/county_population.py

@@ -51,7 +51,7 @@ def get_county_populations_2000s(data_path: Path, data_repo_path: Path) -> pd.Da
             f"https://www2.census.gov/programs-surveys/popest/tables/2000-2010/"
             f"intercensal/county/co-est00int-01-{state.fips}.csv",
         )
-        for state in us.STATES_AND_TERRITORIES
+        for state in us.STATES_AND_TERRITORIES + [us.states.DC]
         if state.fips not in ["60", "66", "69", "72", "78"]  # exclude territories
     ]
 
@@ -83,7 +83,11 @@ def get_county_populations_2000s(data_path: Path, data_repo_path: Path) -> pd.Da
             engine="python",  # for skipfooter
         )
         df["state_code"] = state_code
-        df["County Name"] = df["County Name"].str.lstrip(".")
+
+        # In these CSV files, the total row looks like "Connecticut",
+        # while the rows for each county look like ".Fairfield County".
+        df = df[df["County Name"].str.startswith(".")]
+        df["County Name"] = df["County Name"].str.removeprefix(".")
 
         dfs.append(df)