Deindexing is usually a manual process. It's only automated in the case of certain scripts and their footprints that Google's algorithms specifically look for (examples, YACG, BANS). This has been around a while and it doesn't have anything to do with 'Panda' so stop the Panda Panic.
The process I've observed is that a site ends up on Google's visual review list for a number of reasons but the most common reasons seem to be newly registered exact match domain names, content and/or link building patterns that match known spamming patterns and flag a site for review and competitor reporting.
Once it's in the manual review list, and you won't know that it is until it's too late, your site is at the mercy of a cubical warrior, most likely in India, who'll take a 30 second or less glance at your site's index page and see if it matches a pattern they've been trained to flag as a 'bad' site. You can have the most outstanding content in the world but if your site matches the visual pattern, it's gone. Likewise, you can have a site with terrible content with an HTML template that looks like it was put together by a drunken monkey using FrontPage 98 and you'll probably pass. However, most (but not all) non-ecommerce sites that don't match a known MFA pattern and don't have any ads on the index page will pass.
Occasionally a site will be bumped up the food chain for a more in depth review by the Mountain View staff where links and actual content are reviewed but this isn't common from what I've seen. For example, if you were suddenly ranking #1 for "credit cards" you might get this kind of scrutiny but you probably won't if you're #1 for "purple antique knitting needles".
well written and said. i have a friend who knows an engineer at Google and he says they ("cubical warriors") are swamped with work.. "cubical warrior" lmao