Meta tag analysis blog post

[aciccarello]

As a followup to #149, I'd like to do some analysis of what metaformats are on people's sites. Would do a combination of manual and automated crawling.

Manual Deep Dives

Would want to look at some common sites specifically to see if/how they use specific OGP types

• YouTube (video)
• Vimeo (video)
• LinkedIn (profiles)
• Mastodon (profiles and posts)
• Instagram (profiles and posts, might be rate limited)
• TikTok (profiles and videos)
• Flickr (photos)
• Wikimedia (photos)
• Wikipedia
• GitHub (profiles and other page types)
• News sites
• Medium
• Dev.to
• StackOverflow

Automated Analysis

Collections to parse

• Top 100 sites (might want to view non-homepages)
• personalsit.es (more hand-crafted variety mapping to profiles)
• Jeremy Keith's recent links on adactio.com (would be good link preview test)

Data to collect

• Are microformats on the page?
• Are OGP tags on the page? Which ones?
• Are twitter card tags on the page? (and are they used without OGP)
• How different are the related fallback tags? (og:description vs twitter:description vs description)
• What media tags are used? Are there multiple? Is there alt text?
• Is author defined? Are there multiple? Is the value the full name or something else?
• Cannonical URL vs og:url
• How common is JSON-LD? How complex? Is it used without OGP?