Google BigQuery to rank GitHub projects by watch events per contributor

gistfile1.txt

WITH
  RepoStats AS (
    SELECT
      repo.name AS repository_name,
      -- Proxies for impact: watches and forks
      COUNTIF(type = 'WatchEvent') AS watch_count,
      COUNTIF(type = 'ForkEvent') AS fork_count,
      -- Proxies for contribution: pr, push, review
      COUNTIF(type = 'PullRequestEvent') as pr_count,
      COUNT(DISTINCT `IF`(type = 'PullRequestEvent', actor.login, NULL)) AS pr_actors,
      COUNTIF(type = 'PushEvent') AS push_event_count,
      COUNT(DISTINCT `IF`(type = 'PushEvent', actor.login, NULL)) AS push_actors,
      COUNTIF(type = 'PullRequestReviewCommentEvent') as review_event_count,
      COUNT(DISTINCT `IF`(type = 'PullRequestReviewCommentEvent', actor.login, NULL)) AS review_actors,
    FROM (
      SELECT * FROM `githubarchive.month.2025*` UNION ALL
      SELECT * FROM `githubarchive.year.2024` 
    )
    WHERE
      type IN ('WatchEvent', 'PushEvent', 'ForkEvent', 'PullRequestEvent', 'PullRequestReviewCommentEvent')
    GROUP BY repository_name
  )
SELECT
  repository_name,
  watch_count,
  fork_count,
  pr_count,
  push_event_count,
  review_event_count,
  pr_actors,
  push_actors,
  review_actors,
  FLOOR(SAFE_DIVIDE(watch_count, push_actors)) AS watch_to_push_actor_ratio,
  FLOOR(SAFE_DIVIDE(watch_count, pr_actors)) AS watch_to_pr_actor_ratio,
  FLOOR(SAFE_DIVIDE(watch_count, review_actors)) AS watch_to_review_actor_ratio,
FROM
  RepoStats
WHERE
  push_actors >= 3 AND pr_actors >= 3
ORDER BY watch_to_pr_actor_ratio DESC
LIMIT 1000;