feat: added pageviews analytics functions (#4336) (#4339)

* feat: added pageviews analytics functions (#4336)

* fix: updated outbound sheets link to breaking change in get_flat_data_df (#4336)

* chore: made analytics percent change calculation consistent in sheets-elements (#4336)

* chore: bumped ga package version (#4336)

* fix: no longer double adjusting for time period length (#4336)

* fix: corrected formula for changes (#4336)

* chore: remove logging (#4336)

analytics/analytics_package/analytics/fields.py

@@ -1,6 +1,7 @@
 # Metric names
 METRIC_EVENT_COUNT = 'eventCount'
 METRIC_TOTAL_USERS = 'totalUsers'
+METRIC_PAGE_VIEW = 'screenPageViews'
 
 # DIMENSIONS
 DIMENSION_PAGE_PATH = {

analytics/analytics_package/analytics/sheets_elements.py

@@ -1,9 +1,11 @@
+import numpy as np
 import pandas as pd
 from .charts import get_data_df
 from .fields import *
 from urllib.parse import urlparse
+import datetime as dt
 
-def get_flat_data_df(analytics_params, metrics, dimensions, remove_matches=None):
+def get_flat_data_df(metrics, dimensions, remove_matches=None, **other_params):
     """
     Get a df from the Analytics API with a flat structure (no multiindex).
 
@@ -22,7 +24,7 @@ def get_flat_data_df(analytics_params, metrics, dimensions, remove_matches=None)
     df = get_data_df(
         metrics,
         [dimension["id"] for dimension in dimensions],
-        **analytics_params,
+        **other_params,
     )
     if remove_matches is not None:
         for i, match in enumerate([dimension["remove_matches"] for dimension in dimensions]):
@@ -46,20 +48,20 @@ def get_outbound_links_df(analytics_params):
     pd.set_option('future.no_silent_downcasting', True)
     # Get the builtin "Click" event
     df_builtin_links = get_flat_data_df(
-        analytics_params,
         [METRIC_EVENT_COUNT, METRIC_TOTAL_USERS],
         [DIMENSION_PAGE_PATH, DIMENSION_BUILTIN_URL, DIMENSION_EVENT_NAME],
-        remove_matches=[None, r"\s*", None]
+        remove_matches=[None, r"\s*", None],
+        **analytics_params,
     ).groupby(
         [DIMENSION_PAGE_PATH["alias"], DIMENSION_BUILTIN_URL["alias"]]
     ).sum().reset_index()
 
     # Get the custom "outbound_link_click" event
     df_custom_links = get_flat_data_df(
-        analytics_params, 
         [METRIC_EVENT_COUNT, METRIC_TOTAL_USERS],
         [DIMENSION_EVENT_NAME, DIMENSION_CUSTOM_URL, DIMENSION_PAGE_PATH], 
         remove_matches=[DIMENSION_EVENT_NAME["remove_matches"], r"\(not set\)", None],
+        **analytics_params, 
     ).groupby(
         [DIMENSION_PAGE_PATH["alias"], DIMENSION_CUSTOM_URL["alias"]]
     ).sum().reset_index()
@@ -117,17 +119,119 @@ def get_outbound_links_change(analytics_params, start_current, end_current, star
         "start_date": start_previous,
         "end_date": end_previous,
     }
-    print(analytics_params_month_2)
     df_current = get_outbound_links_df(analytics_params_month_1).set_index(
         ["Page Path", "Outbound Link", "Hostname"]
     )
     df_previous = get_outbound_links_df(analytics_params_month_2).set_index(
         ["Page Path", "Outbound Link", "Hostname"]
     )
+    total_clicks_percent_change = get_change(
+        df_current["Total Clicks"],
+        df_previous["Total Clicks"],
+        start_current,
+        end_current,
+        start_previous,
+        end_previous
+    )
+    total_users_percent_change = get_change(
+        df_current["Total Users"],
+        df_previous["Total Users"],
+        start_current,
+        end_current,
+        start_previous,
+        end_previous
+    )
+    df_reindexed = df_current.reindex(total_clicks_percent_change.index).fillna(0)
+    df_reindexed["Total Clicks Percent Change"] = total_clicks_percent_change
+    df_reindexed["Total Users Percent Change"] = total_users_percent_change
+    return df_reindexed.sort_values(["Total Clicks", "Total Users"], ascending=False, kind="stable").reset_index()
+
+def get_page_views_df(analytics_params):
+    """
+    Get a DF with page views from the Analytics API.
+
+    :param analytics_params: the parameters for the Analytics API, including authentication and property ids
+    :return: a DataFrame with the page views from the Analytics API
+    """
+    df_response = get_flat_data_df(
+        [METRIC_EVENT_COUNT, METRIC_TOTAL_USERS, METRIC_PAGE_VIEW],
+        [DIMENSION_PAGE_PATH, DIMENSION_EVENT_NAME],
+        dimension_filter="eventName==page_view",
+        **analytics_params,
+    ).rename(
+        columns={
+            DIMENSION_PAGE_PATH["alias"]: "Page Path",
+            METRIC_PAGE_VIEW: "Total Views",
+            METRIC_TOTAL_USERS: "Total Users",
+        }
+    )[["Page Path", "Total Views", "Total Users"]].copy()
+    return df_response
+
+def get_page_views_change(analytics_params, start_current, end_current, start_previous, end_previous):
+    """
+    Get a DF with page views from the Analytics API and a comparison for the prior month
+    :param analytics_params: the parameters for the Analytics API, including authentication and property ids
+    :param start_current: the start date for the current month in the format "YYYY-MM-DD"
+    :param end_current: the end date for the current month
+    :param start_previous: the start date for the previous month
+    :param end_previous: the end date for the previous month
+    """
+    analytics_params_current = {
+        **analytics_params,
+        "start_date": start_current,
+        "end_date": end_current,
+    }
+    analytics_params_previous = {
+        **analytics_params,
+        "start_date": start_previous,
+        "end_date": end_previous,
+    }
+    df_current = get_page_views_df(analytics_params_current).set_index("Page Path")
+    df_previous = get_page_views_df(analytics_params_previous).set_index("Page Path")
     combined_index = df_current.index.union(df_previous.index)
     df_current_reindexed = df_current.reindex(combined_index).fillna(0)
     df_previous_reindexed = df_previous.reindex(combined_index)
-    df_current_reindexed["Total Clicks Percent Change"] = (df_current_reindexed["Total Clicks"] / df_previous_reindexed["Total Clicks"]) - 1
-    df_current_reindexed["Total Users Percent Change"] = (df_current_reindexed["Total Users"] / df_previous_reindexed["Total Users"]) - 1
-    return df_current_reindexed.sort_values(["Total Clicks", "Total Users"], ascending=False, kind="stable").reset_index()
-
+    views_percent_change = get_change(
+        df_current_reindexed["Total Views"],
+        df_previous_reindexed["Total Views"],
+        start_current,
+        end_current,
+        start_previous,
+        end_previous,
+    )
+    users_percent_change = get_change(
+        df_current_reindexed["Total Users"],
+        df_previous_reindexed["Total Users"],
+        start_current,
+        end_current,
+        start_previous,
+        end_previous,
+    )
+    df_reindexed = df_current.reindex(views_percent_change.index).fillna(0)
+    df_reindexed["Total Views Percent Change"] = views_percent_change
+    df_reindexed["Total Users Percent Change"] = users_percent_change
+    return df_reindexed.sort_values(["Total Views", "Total Users"], ascending=False, kind="stable").reset_index()
+
+def get_change(series_current, series_previous, start_current, end_current, start_previous, end_previous, combined_index = None):
+    """
+    Get the percent change between two serieses, accounting for different numbers of days in the month.
+    :param series_current: the series representing the current month
+    :param series_previous: the series representing the prior month
+    :param start_current: the start date for the current month in the format "YYYY-MM-DD"
+    :param end_current: the end date for the current month
+    :param start_previous: the start date for the prior month
+    :param end_previous: the end date for the prior month
+    :return: a Series with the change between the two serieses
+    """
+    # Check that both serieses have the same index names
+    assert series_current.index.names == series_previous.index.names
+    # Reindex both serieses to have the same index
+    combined_index = series_current.index.union(series_previous.index)
+    current_length = float((dt.datetime.fromisoformat(end_current) - dt.datetime.fromisoformat(start_current)).days + 1)
+    previous_length = float((dt.datetime.fromisoformat(end_previous) - dt.datetime.fromisoformat(start_previous)).days + 1)
+    assert current_length != 0 and previous_length != 0
+    series_current_reindexed = series_current.reindex(combined_index).fillna(0)
+    # Adjust the values from the prior series to account for the different number of days in the month
+    series_previous_reindexed = (series_previous.reindex(combined_index) * current_length / previous_length)
+    change = ((series_current_reindexed / series_previous_reindexed) - 1).replace({np.inf: np.nan})
+    return change

analytics/analytics_package/setup.py

@@ -2,7 +2,7 @@
 
 setup(
 	name="analytics",
-	version="3.2.0",
+	version="3.3.0",
 	packages=["analytics"],
 	install_requires=["matplotlib", "pandas", "numpy", "google-auth-oauthlib", "google-api-python-client", "gspread", "gspread-formatting"],
 )