In this bounty I'm trying to uncover what activities staking account owners undertake (including, but not limited to staking). First, I tallied the addresses that authorised a staking account to either a 'delegate' or 'deactivate' a stake in 2022, I'll be considering these as stakers.

Second, I looked at transactions in the past month with these addresses as senders and grouped transactions by the program_ID they had to derive the transaction count. I joined this list with flipside's labels table to annotate the program addresses, however I found that some of the most frequently used addresses were without labels. I attempted to identify the labels missing from the top10 using Solscan relying on the address's public name and the log data of it's transactions to guess it's purpose. This is the resulting table of my inquiry:

A large proportion of the calls were made to two addresses I manually annotated as oracles, the Pyth and the Switchboard oracles. These are services typically used by DEX-es and institutions to real-time monitor prices. The second most used app was serum and mango-markets, hence most of the non-oracle traffic went to farming and swapping through dex-es.

The problem with this sort of approach, though is that it cannot distinguish between 'institutional' and 'retail' stakers and the former generate an inordinate amount of traffic through their automated operations, swamping the rest of the signal.