create new table to track weekly unlabeled routes for payments (#2686)

* create new weekly unlabeled routes tracking table, add 'monthly' to title of original table * update ymls
cal-itp · Jun 6, 2023 · 328842c · 328842c
1 parent 0ee2e9c
commit 328842c
Show file tree

Hide file tree

Showing 3 changed files with 84 additions and 3 deletions.
diff --git a/warehouse/models/mart/payments/reliability/_payments_views_tests.yml b/warehouse/models/mart/payments/reliability/_payments_views_tests.yml
@@ -31,7 +31,7 @@ models:
         description: A participant's total ridership on a particular day
       - name: relative_difference
         description: The relative change in a participant's ridership count as compared to the previous day
-  - name: v2_payments_reliability_unlabeled_routes
+  - name: v2_payments_reliability_monthly_unlabeled_routes
     tests:
       - dbt_utils.expression_is_true:
           expression: "n_all_rides >= (n_route_z_rides + n_null_rides)"
@@ -52,3 +52,24 @@ models:
         description: The percentage of unlabeled rides compared to the total number of rides for the given month
       - name: recency_rank
         description: Used to identify a month's recency (for help with filtering)
+  - name: v2_payments_reliability_weekly_unlabeled_routes
+    tests:
+      - dbt_utils.expression_is_true:
+          expression: "n_all_rides >= (n_route_z_rides + n_null_rides)"
+    columns:
+      - name: participant_id
+        description: Littlepay-assigned Participant ID
+      - name: week_start
+        description: The first day of the week for which the count aggregation is being caclulated
+      - name: n_route_z_rides
+        description: The total number of rides labeled as `Route Z` for the given month
+      - name: n_null_rides
+        description: The total number of rides labeled as `Null` for the given month
+      - name: total_unlabeled_rides
+        description: The total number of rides labeled as either `Route Z` or `Null` for the given month
+      - name: n_all_rides
+        description: The total number of rides for the given month
+      - name: pct_unlabeled_rides_to_total_rides
+        description: The percentage of unlabeled rides compared to the total number of rides for the given month
+      - name: recency_rank
+        description: Used to identify a month's recency (for help with filtering)
diff --git a/...payments_reliability_unlabeled_routes.sql → ..._reliability_monthly_unlabeled_routes.sql b/...payments_reliability_unlabeled_routes.sql → ..._reliability_monthly_unlabeled_routes.sql
@@ -41,7 +41,7 @@ aggregations_and_date_spine AS (
         USING (participant_id, month_start)
 ),
 
-v2_payments_reliability_unlabeled_routes AS (
+v2_payments_reliability_monthly_unlabeled_routes AS (
     SELECT
 
         participant_id,
@@ -57,4 +57,4 @@ v2_payments_reliability_unlabeled_routes AS (
     FROM aggregations_and_date_spine
 )
 
-SELECT * FROM v2_payments_reliability_unlabeled_routes
+SELECT * FROM v2_payments_reliability_monthly_unlabeled_routes
diff --git a/...ouse/models/mart/payments/reliability/v2_payments_reliability_weekly_unlabeled_routes.sql b/...ouse/models/mart/payments/reliability/v2_payments_reliability_weekly_unlabeled_routes.sql
@@ -0,0 +1,60 @@
+WITH payments_rides AS (
+    SELECT * FROM {{ ref('fct_payments_rides_v2') }}
+),
+
+payments_tests_weekly_date_spine AS (
+    SELECT * FROM {{ ref('payments_tests_weekly_date_spine') }}
+),
+
+count_rides AS (
+    SELECT
+
+        spine.participant_id,
+        spine.week_start,
+        COUNTIF(route_id = 'Route Z') AS n_route_z_rides,
+        COUNTIF(route_id IS NULL) AS n_null_rides,
+        COUNTIF(route_id = '') AS n_empty_string_rides,
+        COUNT(*) AS n_all_rides
+
+    FROM payments_tests_weekly_date_spine AS spine
+    INNER JOIN payments_rides
+        ON (spine.participant_id = payments_rides.participant_id) AND transaction_date_pacific >= week_start AND transaction_date_pacific <= week_end
+    GROUP BY week_start, participant_id
+),
+
+aggregations_and_date_spine AS (
+    SELECT
+
+        date_spine.participant_id,
+        date_spine.week_start,
+        count_rides.n_route_z_rides,
+        count_rides.n_null_rides,
+        count_rides.n_empty_string_rides,
+        count_rides.n_all_rides,
+
+        (count_rides.n_route_z_rides + count_rides.n_null_rides + count_rides.n_empty_string_rides) AS total_unlabeled_rides,
+
+        SAFE_DIVIDE((count_rides.n_route_z_rides + count_rides.n_null_rides + count_rides.n_empty_string_rides), count_rides.n_all_rides) * 100 AS pct_unlabeled_rides_to_total_rides
+
+    FROM payments_tests_weekly_date_spine AS date_spine
+    LEFT JOIN count_rides
+        USING (participant_id, week_start)
+),
+
+v2_payments_reliability_weekly_unlabeled_routes AS (
+    SELECT
+
+        participant_id,
+        week_start,
+        n_route_z_rides,
+        n_null_rides,
+        total_unlabeled_rides,
+        n_empty_string_rides,
+        n_all_rides,
+        pct_unlabeled_rides_to_total_rides,
+        RANK() OVER (PARTITION BY participant_id ORDER BY week_start DESC) AS recency_rank
+
+    FROM aggregations_and_date_spine
+)
+
+SELECT * FROM v2_payments_reliability_weekly_unlabeled_routes