This is just a short post to write down an idea (that I’m assuming others have had) that for me was born of frustration on the Documenting the Now project and inspiration while reading Escobar (2017), here on the 5th anniversary of the killing of Mike Brown.
Let’s get the frustration out of the way first.
We’re trying to create tools that help people document events that get shared in social media. We started out trying to collect what was happening in Twitter during the Ferguson Uprising. As we all learned later with Cambridge Analytica it was actually a bit too easy to get at this data, especially if you have the technical skills to figure out how to talk to their API, and the credibility to get a set of app keys.
Getting the data is the easy part. Getting the consent of the people that created it, so that you can archive it is something else entirely.
Thankfully now it’s harder for just anyone to get at this data. But the policies around who gets access and how they get access to it act as a gateway that limits the people and organizations who can use the data. This is by design. Increasingly this is a big problem for researchers who are no longer able to study social media as they could before Cambridge Analytica (Bruns, 2019 ; Freelon, 2018). But its also a huge problem for just regular people, the billions of people who use social media every day.
However (and this is where the inspiration comes in) through a series of workshops starting in Ferguson and now being held in locations around the US organized by the tireless Bergis Jules, the Documenting the Now team has had the opportunity to hear from community and activist groups who are using social media as a place to document their own work. And they would like to get their data out to have it for themselves.
For example the Texas After Violence Project we met with last week have a website, and a Twitter account, and an Instagram account. They use these spaces to publish oral histories that help document state-sanctioned violence on people and communities. While there are tools for pushing content into these spaces there aren’t really any for getting their stuff out again. They want to have their data in case the platform decides their content is no longer welcome there, or the platforms policies change and they no longer want their content on there. As you can imagine there are lots of reasons why people might want to get their own data out of a social media platform. At its heart its just the nagging awareness that lots of copies keeps stuff safe.
And yet (back to the frustration) it’s a big investment for these groups to be able to download and install the DocNow application as it stands now. It involves standing up an application in the cloud that’s composed of a database or two, and a webserver, and managing little worker processes that harvest metadata, unshorten links, etc. Even with devops tools to automate that build process it’s complicated, and kinda expensive for an activist organization.
Plus the DocNow app is centered on Twitter, because of the public API access it still provide, and its use by journalists and the media. But the world of social media is so much bigger that just Twitter. Even accessing Twitter requires someone to create an app and get approved to get the keys to run it, which includes convincing Twitter that you are worthy of that, which is an increasingly opaque process to navigate. We have a command line tool twarc which doesn’t require such an expensive set up as DocNow, but the command line isn’t for everyone, and it still requires you to get app keys. Then there are tools like twint that scrape content out of Twitter because they can, even if that’s against the platform Terms of Service.
A year ago the General Data Protection Regulation (GDPR) kicked in in the EU, and a nice side effect of Article 15 is that social media platforms had to provide a way for you to get your data out of their platform. Alexandra Dolan-Mescal recently created a short zine called Social Control the provides people with instructions of how to get their data out of Facebook, Twitter, Instagram and Snapchat. Some of the platforms provide rudimentary guides to the downloaded data, but that’s not always the case.
So what’s the idea already you say?
What if there were an application that you could run on your computer that allowed you to import these GDPR archives and access your photos, posts, etc in a meaningful way so you could have them yourself? This idea has been tried before by folks like Aaron Straup-Cope with his PrivateSquare, or Dan Phiffer’s smol.archive. But these have been a bit hamstrung by the fact that they require you to get keys so that you can access your data. This, and setting up an application aren’t things just anybody can typically do. Working off of the GDPR download would be slower, and less up to date, but it would at least would only require you to download a zip file from the said platform and then start up a desktop app and import it. Then you’d have to repeat that process occasionally.
I mentioned this idea on Twitter, and got some interesting responses. One of which is to try out Solid which I should look at more closely. I’m much more familiar with Mastodon and Scuttlebutt in the federated/distributed social media space, and understood Solid to be working to achieve similar goals. I absolutely think we need these new models for social media, but we also have a lot of people using centralized forms that have evolved on the (once) distributed web. We have an obligation to help people work with the web we have now, while also helping build a better one.
I'm starting to wonder if the only way for people to manage their own social media data is for them to periodically download their GDPR archive and give it to an app that they can run themselves (off cloud). Has this been done already? Is it a crazy idea?— Ed Summers ((???)) August 9, 2019
Another alternative to a new app would be to get folks working with the soon to be released Webrecorder-Desktop. We actually did this in our workshop with the Texas After Violence Project. Webrecorder-Desktop is an application you run on your workstation that you can use to archive and replay all kinds of web content. Since it uses your browser to do it it creates a high fidelity copy that is driven by how you interact with content in the browser. One thing that could be tedious is having to do repetitive actions like click on the detail of an Instagram post to get the larger image and comments. Fortunately they are working on what they called Behaviors that allow you to semi-automate the collection, so you could point your Webrecorder-Desktop app at your Instagram account and tell it to autoscroll the page and click on each item to get the details automatically.
So maybe instead of creating a new app to make sense of the GDPR download, perhaps it would be better (and more collaborative) to think about enhancing these behaviors to allow people to collect content from these sites? Maybe Webrecorder-Desktop could use the GDPR archive in some way in order to know what to crawl?
The obvious benefit to using the GDPR data is that the only person who has access to it is you. You are the only person who can request it and download it. Services like TwArxiv that ask you to upload your GDPR data to another service on the web are interesting experiments, but they are also risky propositions that you have every right to be wary of. Plus, you may not want to analyze your data, but just have your own copy of it, for yourself. Because it’s yours, right?
PS. I don’t particularly want to argue about whether the GDPR download is an archive or not. We all know that it’s not that kind of archive. But it’s a useful convenience to call it that. And maybe if it were more useful, and could be managed and backed up, it would be a little bit more like that kind of archive ;-)
Bruns, A. (2019). After the `APIcalypse’: social media platforms and their fight against critical scholarly research. Information, Communication & Society.
Escobar, A. (2017). Designs for the Pluriverse. Duke University Press.
Freelon, D. (2018). Computational research in the post-API age. Political Communication. Retrieved from https://osf.io/preprints/socarxiv/56f4q