Ever since my discovery of the IndieWeb movement, I've wanted to free myself from Facebook (and Instagram) and their brand of surveillance capitalism. I want to own my own data, and be in control of how it is shared, and I don't want it to be used for advertising.
I've had this incarnation of a personal website for a few years, and have mostly been following the POSSE publishing model, publishing most forms of content on my website, and then automatically (or manually) syndicating that content to silos like Facebook and Twitter. But, much of my content still remains trapped inside of Facebook and Instagram.
As of March 4, 2018, I've pulled the vast majority of my Facebook content into my website, and all of my Instagram photos into my website, paving the way for me to delete myself from Facebook (and potentially Instagram) by the end of 2018. What follows is a high-level overview of how I made the move.
Exporting Data from Facebook
While Facebook does offer an export feature, its extremely limited, only includes very low resolution versions of your photos, and is generally very difficult to process programmatically. After some research, I discovered the excellent fb-export project on GitHub. Once installed, this tool will dump a huge amount (though, not quite all) of your Facebook data into machine-readable JSON files.
Since my website is compatible with the Micropub publishing standard, I then needed to convert this Facebook-native JSON data into microformats2 formatted JSON. Enter granary, an amazing swiss-army knife of IndieWeb by Ryan Barrett. Using granary, I whipped up a quick script that transforms the exported data into native microformats2 formatted JSON:
Publishing Liberated Data
At this point, I had a directory full of data ready to publish. Sort of. Unfortunately, not all of the data is easily translatable, or even desirable, to publish to my website. As a result, I created another script that let me, on a case by case basis, publish a piece of content, choose to skip it entirely, or save it to deal with later.
After running this script, I had a significant amount of my data copied from Facebook to my website. Huzzah!
Dealing with Photo Albums
Facebook has a "photo albums" feature, and I definitely wanted to get those memories onto my website. Again, I wrote a script that processes the exported data, and selectively allows me to upload all of the photos in an album to my website via Micropub, and then drops microformats2 JSON out that I could publish later.
Once I finished processing and uploading all of the photos for the albums I wished to copy over, I ran a simple utility script I keep around to publish all of the albums as new posts to my website.
Here are some of the results:
cleverdevil.io/ 2015/ the-lunds-visit-for-easter-and-palm-springs
cleverdevil.io/ 2014/ maui-2014
cleverdevil.io/ 2012/ hawaii-2012
cleverdevil.io/ 2012/ out-and-about-with-colette
cleverdevil.io/ 2009/ europe-trip---part-1
Notice, one of these comes all the way back from 2009!
There are still quite a few photos and other types of posts that I haven't yet been able to figure out how to migrate. Notably, Facebook has strange special albums such as "iOS Uploads," "Mobile Uploads," and "iPhoto Uploads" that represent how the photos were uploaded, not so much a group of related photos. Unfortunately, the data contained in the export produced by fb-export isn't quite adequate to deal with these yet.
Still, I am quite pleased with my progress so far. Time to move on to Instagram!
Instagram has been slowly deteriorating as a service for years, so much so that I decided to completely stop publishing to Instagram earlier this year. It turns out, dealing with Instagram is a lot easier than Facebook when it comes to liberating your data.
Downloading My Data
After some research, I found instaLooter on GitHub, which allowed me to quickly export every single photo in its original resolution, along with nearly every bit of data I needed... except the photo captions. I ran instaLooter, and embedded the unique identifier in the filenames (which instaLooter refers to as the "code').
Getting Metadata and Publishing
I wrote a script that used granary to lookup the photo metadata and publish to my website via Micropub:
Note, I used the non-JSON form of Micropub in this case, because Known's Micropub implementation doesn't properly handle JSON for photos yet.
It turns out, that with a little knowhow, and a lot of persistence, you can liberate much of your data from Facebook and Instagram. I feel well on target to my goal of leaving Facebook (and maybe Instagram) entirely.