Yes, absolutely. For context, we do stock market analysis but we approach it from the perspective of stock prices being entirely supply and demand, and we're looking for stocks which are about to receive a lot of new demand. We handle A LOT of data. Well, a lot for what we do, we're not exabyte scale yet.
We use Postmodern for Postgres interaction. Have no complaints at all. You're able to use s-sql which is SQL statements as s-exps but can drop into just passing it strings if required.
Dexador is how we're doing web scraping (and requests should we need to). Top tier, no complaints. Exactly what you want from it. Plump and lquery are also used for parsing HTML
Jonathan is where I'll fight with most people in the CL community. It's the best JSON library. When I say fight, I mean it, I'll throw hands at anyone recommending cl-json or jsown.
Local-time is something we depend upon a lot. Great for timestamp manipulation.
cl-cron for scheduling. I like it well enough.
I'm also a big fan of Hunchentoot for the web requests. All static pages (currently all addressable webpages) are generated by Spinneret.
There's a few more libraries but they may be more niche or more specific to us.