a bad "page" could be:

• never or only rarely visited (number)
• visited only for a very short time in average or
• visited only for a very short time summed up in total

As long as it's possible to remove/diable the feature. If you "only" have 500 pages it's not too bad I presume but I got 11 000 different web pages and maybe 4000 PDF so yes, we have many "unviewed" or low-visited pages over time...

of course it should be only optional - maybe even as an plugin.
In addition it may make sense to have a possibility to restrict (white or blacklist) a path within url...

as first approach crawling could be easily done by using a standard (google-)sitemap.
These are available on most websites and contains the urls of all pages in a standardized xml format

<?xml version="1.0" encoding="UTF-8"?>
<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">
  <url>
    <loc>http://www.example.net/outsite/black-cats.html</loc>
    <lastmod>2015-09-22</lastmod>
    <changefreq>monthly</changefreq>
    <priority>0.8</priority>
  </url>
  <url>
    ....
  </url>
</urlset>

For this one only need a setting for the url of it (most times it's http://www.example.net/sitemap.xml)

Since the comparison should only be done with very low frequency (e.g.once a month) or on request (e.g. via Button click to schedule it to make it next night) there shouldn't be a big performance problem at all...

In addition a setting for comparison period is needed:
e.g. (not)visited within last month, last year, for all data...
=> hmm maybe this could be directly set by using segment editor?

could e.g. be part of a new plugin which provides a new menu point: "Find the bad"
which contains the functions:
"Invers" page tracking - to find bad pages #10519
"Invers" Custom Event tracking - to find bad elements/functions #5186
"Invers" Content tracking - to find bad pieces of content #10520