Donk_enby had originally intended to grab data only from the day of the Capitol takeover, but found that the poor construction and security of Parler allowed her to capture, essentially, the entire website. That ended up being 56.7 terabytes of data, which included every public post on Parler, 412 million files in all—including 150 million photos and more than 1 million videos. Each of these had embedded metadata like date, time and GPS coordinates—unlike most social media sites, Parler does not strip metadata from media its users upload, which, crucially, could be useful for law enforcement and open source investigators.
I’m no SysAdmin but surely one of the first things you do if you run a web service which lets users upload media is to make sure that you strip the EXIF data. It’s hilarious that Parler didn’t do this.
But the vast majority of video on Parler had nothing to do with the riot. So ProPublica have done the legwork [c] and released an archive of around 500 videos from the riot itself. All neatly organised too (unlike the rioters).
You can view them all here. It’s fascinating stuff.