i AS (
SELECT COUNT(*) AS overlap_count, the_geom,
  row_number() OVER (ORDER BY COUNT(*)) AS overlap_group
FROM h
GROUP BY the_geom
)