Тестовый коммит

Reviewed-on: https://gitea.de-infra.servehttp.com/de-infrastructure/DWH/pulls/114

aharlamcev/PET-project_task3

@@ -0,0 +1,26 @@
+CREATE MATERIALIZED VIEW platform_summary AS
+with users_counts as (
+					select 
+					COUNT(distinct u.id) as total_users,
+					COUNT(distinct lv.user_id) as users_with_views
+					from users u
+					left join lesson_views lv on lv.user_id = u.id
+					),
+courses_lessons_count as (
+					select COUNT(distinct c.id) as total_courses,
+					COUNT(distinct l.id) as total_lessons
+					from courses c
+					left join lessons l on l.course_id = c.id
+					),
+views_count as (
+				select COUNT(distinct lv.lesson_id) as lessons_with_views,
+				COUNT(distinct lv.id) as total_views
+				from lesson_views lv
+				)
+select uc.total_users, clc.total_courses, clc.total_lessons, uc.users_with_views,
+ROUND(1.00 * clc.total_lessons / nullif(clc.total_courses, 0), 2) as avg_lessons_per_course,
+ROUND(1.00 * vc.total_views / nullif(clc.total_lessons, 0), 2) as avg_views_per_lesson
+from users_counts uc
+cross join courses_lessons_count clc
+cross join views_count vc
+;

aharlamcev/PET-project_task4

@@ -0,0 +1,28 @@
+CREATE MATERIALIZED VIEW users_cohort_analysis AS
+with needed_users as (
+						select distinct u.id as user_id, 
+						u.registration_date as reg_date, e.course_id, lv.id as view_id, 
+						lv.lesson_id as viewed_lesson_id, 
+						FIRST_VALUE(lv.viewed_at) over (partition by lv.user_id order by lv.viewed_at) as first_view_date
+						from users u
+						left join enrollments e on u.id = e.user_id
+						left join lesson_views lv on e.user_id = lv.user_id
+						order by user_id
+						),
+CTE as (
+		select 
+		TO_CHAR(reg_date, 'YYYY-MM') as cohort_month, 
+		COUNT(distinct user_id) as users_in_cohort,
+		COUNT(distinct case when user_id in (select user_id from enrollments) then user_id end) as enrolled_users,
+		COUNT(distinct case when user_id in (select user_id from lesson_views) then user_id end) as active_users,
+		COUNT(course_id) as courses_count,
+		extract('days' from AVG(first_view_date - reg_date)) as avg_time_to_first_view
+		from needed_users
+		group by cohort_month
+		)
+select cohort_month, users_in_cohort, enrolled_users, active_users,
+ROUND(100.0 * active_users / users_in_cohort, 2) as retention_rate,
+courses_count::numeric / users_in_cohort as avg_courses_per_user,
+avg_time_to_first_view
+from CTE
+;

aharlamcev/postgresql/Pet-projectsql.sql

@@ -283,7 +283,7 @@ from total_counts
 ;
 
 -----------------------------------------------------------------------------------------------------
--- 4 Задание Когортный анализ пользователей
+-- 4 Задание Когортный анализ пользователей (Для каждого месяца регистрации)
 create materialized view user_cohort_analysis AS
 with needed_users as (
 						select distinct u.id as user_id, u.registration_date as reg_date, e.course_id, lv.id as view_id, 
@@ -305,7 +305,6 @@ CTE as (
 		COUNT(course_id) as courses_count,
 		extract('days' from AVG(first_view_date - reg_date)) as avg_time_to_first_view
 		from needed_users
-		where date_trunc('month', reg_date)::date = '2023-03-01'
 		group by cohort_month
 		)
 select cohort_month, users_in_cohort, enrolled_users, active_users,
@@ -340,7 +339,7 @@ views_cte as (
 select lic.lesson_id, lic.lesson_title, lic.course_title, 
 vc.total_views, 
 vc.unique_viewers, 
-coalesce(vc.avg_views_per_user, 0),
+coalesce(vc.avg_views_per_user, 0) as avg_views_per_user,
 dense_rank() OVER (ORDER BY vc.total_views desc) as popularity_rank,
 ROUND(100.0 * unique_viewers / enrolled_to_course, 2) as completion_rate
 from lessons_in_courses lic
@@ -441,5 +440,27 @@ funnel_stages as (
 ;
 -- Признаюсь 6-е задание сделал с помощью нейросети, потому что сам вообще никак к нему подобраться не мог.
 
-
-
+-----------------------------------------------------------------------------------------------------------------
+-- Доп. Задание №1 Анализ drop-off точек
+select course_id, stage_name, users_count,
+100 - conversion_from_previous as lost_precentage
+from conversion_funnel
+where stage_name = 'started'
+order by lost_precentage desc
+;
+--------------------------------------------------------------------------------------------------------------------
+-- Доп. Задание №2 Топ-3 урока с повторными просмотрами
+select lesson_id
+from lesson_popularity_ranking
+where avg_views_per_user > 1
+limit 3
+;
+---------------------------------------------------------------------------------------------------------------------
+-- Доп. Задание №3 Лучшая когорта по retention
+select cohort_month, retention_rate
+from user_cohort_analysis
+order by retention_rate desc
+limit 2
+;
+-- В нашем случае два месяца с одинаковыми лучшими показателями retention это 2023-01 и 2023-02. Это является следствием того, что у нас
+-- наибольшее количество зарегистрированных пользователей именно в эти месяцы, а именно 2023-01 = 40%, а 2023-02 = 20%, соотвественно 60% пользователей были зарегистрированы в эти месяцы.

test.txt

@@ -1 +1 @@
-test коммита в Gitea
+Hello from Python branch

test/postgresql/homework.sql

@@ -0,0 +1 @@
+select * from users;

test/postgresql/homework2.sql

@@ -0,0 +1 @@
+select * from users;