Merge pull request #1505 from cmu-delphi/ndefries/add-hosp-prop

Add hospital admissions prop signal to CPR

Author: krivard

ansible/templates/dsew_community_profile-params-prod.json.j2

@@ -26,7 +26,9 @@
       "ref_window_size": 7,
       "smoothed_signals": [
         "naats_total_7dav",
-        "naats_positivity_7dav"
+        "naats_positivity_7dav",
+        "confirmed_admissions_covid_1d_prop_7dav",
+        "confirmed_admissions_covid_1d_7dav"
       ]
     }
   }

dsew_community_profile/delphi_dsew_community_profile/constants.py

@@ -50,22 +50,34 @@ class Transform:
 SIGNALS = {
     "total": {
         "is_rate" : False,
-        "api_name": "naats_total_7dav"
+        "api_name": "naats_total_7dav",
+        "make_prop": False
     },
     "positivity": {
         "is_rate" : True,
-        "api_name": "naats_positivity_7dav"
+        "api_name": "naats_positivity_7dav",
+        "make_prop": False
     },
     "confirmed covid-19 admissions": {
         "is_rate" : False,
-        "api_name": "confirmed_admissions_covid_1d_7dav"
+        "api_name": "confirmed_admissions_covid_1d_7dav",
+        "make_prop": True,
+        "api_prop_name": "confirmed_admissions_covid_1d_prop_7dav"
     }
 }
 
 COUNTS_7D_SIGNALS = {key for key, value in SIGNALS.items() if not value["is_rate"]}
 
-def make_signal_name(key):
-    """Convert a signal key to the corresponding signal name for the API."""
+def make_signal_name(key, is_prop=False):
+    """Convert a signal key to the corresponding signal name for the API.
+
+    Note, this function gets called twice with the same `key` for signals that support
+    population-proportion ("prop") variants.
+    """
+    if is_prop:
+        return SIGNALS[key]["api_prop_name"]
     return SIGNALS[key]["api_name"]
 
-NEWLINE="\n"
+NEWLINE = "\n"
+IS_PROP = True
+NOT_PROP = False

dsew_community_profile/delphi_dsew_community_profile/pull.py

@@ -11,8 +11,9 @@
 
 from delphi_utils.geomap import GeoMapper
 
-from .constants import TRANSFORMS, SIGNALS, COUNTS_7D_SIGNALS, NEWLINE
-from .constants import DOWNLOAD_ATTACHMENT, DOWNLOAD_LISTING
+from .constants import (TRANSFORMS, SIGNALS, COUNTS_7D_SIGNALS, NEWLINE,
+    IS_PROP, NOT_PROP,
+    DOWNLOAD_ATTACHMENT, DOWNLOAD_LISTING)
 
 # YYYYMMDD
 # example: "Community Profile Report 20211104.xlsx"
@@ -248,7 +249,7 @@ def _parse_sheet(self, sheet):
             if (sheet.level == "msa" or sheet.level == "county") \
                 and self.publish_date < datetime.date(2021, 1, 8) \
                 and sig == "confirmed covid-19 admissions":
-                self.dfs[(sheet.level, sig)] = pd.DataFrame(
+                self.dfs[(sheet.level, sig, NOT_PROP)] = pd.DataFrame(
                         columns = ["geo_id", "timestamp", "val", \
                             "se", "sample_size", "publish_date"]
                     )
@@ -258,7 +259,7 @@ def _parse_sheet(self, sheet):
             assert len(sig_select) > 0, \
                 f"No {sig} in any of {select}\n\nAll headers:\n{NEWLINE.join(list(df.columns))}"
 
-            self.dfs[(sheet.level, sig)] = pd.concat([
+            self.dfs[(sheet.level, sig, NOT_PROP)] = pd.concat([
                 pd.DataFrame({
                     "geo_id": sheet.geo_id_select(df).apply(sheet.geo_id_apply),
                     "timestamp": pd.to_datetime(self.times[si[0]][sig]),
@@ -271,7 +272,7 @@ def _parse_sheet(self, sheet):
             ])
 
         for sig in COUNTS_7D_SIGNALS:
-            self.dfs[(sheet.level, sig)]["val"] /= 7 # 7-day total -> 7-day average
+            self.dfs[(sheet.level, sig, NOT_PROP)]["val"] /= 7 # 7-day total -> 7-day average
 
 
 def as_cached_filename(params, config):
@@ -390,13 +391,46 @@ def fetch_new_reports(params, logger=None):
     # add nation from state
     geomapper = GeoMapper()
     for sig in SIGNALS:
-        state_key = ("state", sig)
+        state_key = ("state", sig, NOT_PROP)
         if state_key not in ret:
             continue
-        ret[("nation", sig)] = nation_from_state(
+        ret[("nation", sig, NOT_PROP)] = nation_from_state(
             ret[state_key].rename(columns={"geo_id": "state_id"}),
             sig,
             geomapper
         )
 
+    for key, df in ret.copy().items():
+        (geo, sig, _) = key
+        if SIGNALS[sig]["make_prop"]:
+            ret[(geo, sig, IS_PROP)] = generate_prop_signal(df, geo, geomapper)
+
     return ret
+
+def generate_prop_signal(df, geo, geo_mapper):
+    """Transform base df into a proportion (per 100k population)."""
+    if geo == "state":
+        geo = "state_id"
+    if geo == "county":
+        geo = "fips"
+
+    # Add population data
+    if geo == "msa":
+        map_df = geo_mapper.get_crosswalk("fips", geo)
+        map_df = geo_mapper.add_population_column(
+            map_df, "fips"
+        ).drop(
+            "fips", axis=1
+        ).groupby(
+            geo
+        ).sum(
+        ).reset_index(
+        )
+        df = pd.merge(df, map_df, left_on="geo_id", right_on=geo, how="inner")
+    else:
+        df = geo_mapper.add_population_column(df, geo, geocode_col="geo_id")
+
+    df["val"] = df["val"] / df["population"] * 100000
+    df.drop(["population", geo], axis=1, inplace=True)
+
+    return df

dsew_community_profile/delphi_dsew_community_profile/run.py

@@ -58,14 +58,14 @@ def replace_date_param(p):
     run_stats = []
     dfs = fetch_new_reports(params, logger)
     for key, df in dfs.items():
-        (geo, sig) = key
+        (geo, sig, is_prop) = key
         if sig not in params["indicator"]["export_signals"]:
             continue
         dates = create_export_csv(
             df,
             params['common']['export_dir'],
             geo,
-            make_signal_name(sig),
+            make_signal_name(sig, is_prop),
             **export_params
         )
         if len(dates)>0:

dsew_community_profile/params.json.template

@@ -32,7 +32,9 @@
       "ref_window_size": 7,
       "smoothed_signals": [
         "naats_total_7dav",
-        "naats_positivity_7dav"
+        "naats_positivity_7dav",
+        "confirmed_admissions_covid_1d_prop_7dav",
+        "confirmed_admissions_covid_1d_7dav"
       ]
     }
   }

dsew_community_profile/tests/params.json.template

@@ -25,7 +25,9 @@
       "ref_window_size": 7,
       "smoothed_signals": [
         "naats_total_7dav",
-        "naats_positivity_7dav"
+        "naats_positivity_7dav",
+        "confirmed_admissions_covid_1d_prop_7dav",
+        "confirmed_admissions_covid_1d_7dav"
       ]
     }
   }

dsew_community_profile/tests/test_pull.py

@@ -9,7 +9,7 @@
 
 from delphi_dsew_community_profile.pull import DatasetTimes
 from delphi_dsew_community_profile.pull import Dataset
-from delphi_dsew_community_profile.pull import fetch_listing, nation_from_state
+from delphi_dsew_community_profile.pull import fetch_listing, nation_from_state, generate_prop_signal
 
 example = namedtuple("example", "given expected")
 
@@ -213,3 +213,84 @@ def test_nation_from_state(self):
                 'sample_size': [None],}),
             check_like=True
         )
+
+    def test_generate_prop_signal_msa(self):
+        geomapper = GeoMapper()
+        county_pop = geomapper.get_crosswalk("fips", "pop")
+        county_msa = geomapper.get_crosswalk("fips", "msa")
+        msa_pop = county_pop.merge(county_msa, on="fips", how="inner").groupby("msa").sum().reset_index()
+
+        test_df = pd.DataFrame({
+                'geo_id': ['35620', '31080'],
+                'timestamp': [datetime(year=2020, month=1, day=1)]*2,
+                'val': [15., 150.],
+                'se': [None, None],
+                'sample_size': [None, None],})
+
+        nyc_pop = int(msa_pop.loc[msa_pop.msa == "35620", "pop"])
+        la_pop = int(msa_pop.loc[msa_pop.msa == "31080", "pop"])
+
+        expected_df = pd.DataFrame({
+                'geo_id': ['35620', '31080'],
+                'timestamp': [datetime(year=2020, month=1, day=1)]*2,
+                'val': [15. / nyc_pop * 100000, 150. / la_pop * 100000],
+                'se': [None, None],
+                'sample_size': [None, None],})
+
+        pd.testing.assert_frame_equal(
+            generate_prop_signal(
+                test_df.copy(),
+                "msa",
+                geomapper
+            ),
+            expected_df,
+            check_like=True
+        )
+    def test_generate_prop_signal_non_msa(self):
+        geomapper = GeoMapper()
+
+        geos = {
+            "state": {
+                "code_name": "state_id",
+                "geo_names": ['pa', 'wv']
+            },
+            "county": {
+                "code_name": "fips",
+                "geo_names": ['36061', '06037']
+            },
+            # nation uses the same logic path so no need to test separately
+            "hhs": {
+                "code_name": "hhs",
+                "geo_names": ["1", "4"]
+            }
+        }
+
+        for geo, settings in geos.items():
+            geo_pop = geomapper.get_crosswalk(settings["code_name"], "pop")
+
+            test_df = pd.DataFrame({
+                    'geo_id': settings["geo_names"],
+                    'timestamp': [datetime(year=2020, month=1, day=1)]*2,
+                    'val': [15., 150.],
+                    'se': [None, None],
+                    'sample_size': [None, None],})
+
+            pop1 = int(geo_pop.loc[geo_pop[settings["code_name"]] == settings["geo_names"][0], "pop"])
+            pop2 = int(geo_pop.loc[geo_pop[settings["code_name"]] == settings["geo_names"][1], "pop"])
+
+            expected_df = pd.DataFrame({
+                    'geo_id': settings["geo_names"],
+                    'timestamp': [datetime(year=2020, month=1, day=1)]*2,
+                    'val': [15. / pop1 * 100000, 150. / pop2 * 100000],
+                    'se': [None, None],
+                    'sample_size': [None, None],})
+
+            pd.testing.assert_frame_equal(
+                generate_prop_signal(
+                    test_df.copy(),
+                    geo,
+                    geomapper
+                ),
+                expected_df,
+                check_like=True
+            )