select week, activeness, count(sender) as n_address

from (with um as (select date_trunc('day', t) as week, attribute_value as sender, sum(n_txn) as tot_swap_made

from ethereum.

select date(block_timestamp) as t, attribute_value, count(tx_id) as n_txn

from

where msg_type='token_swapped' and attribute_key='sender'

group by 1,2 order by 1)

group by 1,2 order by 1),

gauge as (

select date_trunc('week', day) as week , avg(txn_per_user) as avg_txn_per_user

from (select date_trunc('day',block_timestamp) as day, count(distinct attribute_value) as n_sender , count(tx_id) as n_txn_swap, (n_txn_swap/n_sender) as txn_per_user --attribute_value,

from osmosis.CORe.fact_msg_attributes

where msg_type='token_swapped' and attribute_key='sender'

group by 1 order by 1

-- order by 4

)

group by 1 order by 1

)

select um.week, um.sender, tot_swap_made, avg_txn_per_user, case when tot_swap_made>=avg_txn_per_user then 'Active User' else 'Inactive User' end as Activeness

from gauge