Question is - what is "high volume"
Maybe set a fixed number of high volume slots and put only the most traded in BTC value in them, then split the rest (middle/junk coins) with an aggregated amount of trade level or/and bids which must be reached to escape the "junk" qualification (I've seen coins with less than 1BTC of bids, they certainly aren't worth mining: by the time the coins would have matured all bids could have gone).
Once the rules are set, there shouldn't be much complaining: this would be centered around the number of slots and level values (which could be dynamically chosen by users).
In the API don't try to split things, just add the relevant information (volume) by creating new keys. API users will always know best how they want to process the data, just feed them all you can and they will be satisfied.
One tip for computing mining efficiency. My observation of orphan rates points to a collision window on a coin P2P network of ~5 seconds. What this means is that the orphan probability on a stable network (with no big hashrate swings) is
5/block_intervalBitcoin should have 5/600 = 0.83% orphans
Worldcoin should have 5/15 = 33% orphans
There's a margin of error here: some networks are a little more stable/fast than others but the numbers don't deviate much from that. When the flock comes on a new coin, the actual block interval is smaller by quite a bit until difficulty is adjusted: the orphan rates rises often to insane levels (that's not easy to detect unless you are solo-mining the coin yourself and I actually coded fall-back measures in my coin-choosing bot to detect such cases).
This is crucial for computing mining efficiency, I've face-palmed more than once when people mined Worldcoin and other fast block interval coins like there was no tomorrow, most of them did probably loose quite a bit of coins while coinchoose.com made them look profitable.