I don’t think aggregating by timeframe would work well. For example, when browsing large relays like relay.damus.io, even aggregating notes from just 24 hours could be very difficult and would definitely hit the relay’s rate limits. On the other hand, for smaller relays, there might be so few notes in 24 hours that aggregation wouldn’t even be necessary. I’d prefer to aggregate by count instead — for example, every 500 or 1000 notes.
