Identify anomalous daily spending by subaccount

Postgres SQL, as run

select (charge_period_start at time zone 'UTC')::date as day,
       case when publisher_name is not null and publisher_name is distinct from invoice_issuer_name then publisher_name else provider_name end as service_provider_name,
       sub_account_id as sub_account_id,
       sum(effective_cost) as daily_effective_cost
from focus.cost_and_usage
where charge_period_start >= $1::timestamptz
  and charge_period_end < $2::timestamptz
group by (charge_period_start at time zone 'UTC')::date, case when publisher_name is not null and publisher_name is distinct from invoice_issuer_name then publisher_name else provider_name end, sub_account_id
order by daily_effective_cost desc nulls last, day asc, service_provider_name asc nulls last, sub_account_id asc nulls last
limit 500

Parameters: $1 = period_start, $2 = period_end

Original FOCUS SQL

SELECT
  DATE(ChargePeriodStart) AS Day,
  ServiceProviderName,
  SubAccountId,
  SUM(EffectiveCost) AS DailyEffectiveCost
FROM focus_data_table
WHERE ChargePeriodStart >= ? and ChargePeriodEnd  < ?
GROUP BY
  DATE(ChargePeriodStart) AS StartDay,
  ServiceProviderName,
  SubAccountId