Monthly Archives: August 2018
3 Comments Posted by Bernhard on August 11th 2018 @ 8:24 am
For the last ten years, I have been developing a Facebook app for researchers called Netvizz. It started as part of a class on advanced web programming I was teaching at the département hypermédia and morphed quite a bit over the years, often in response to API changes. After losing the ability to generate graph files for friendships, it kind of stabilized around analysis for Facebook Pages and – to a lesser degree – Groups. If you check the almost 300 scholarly papers that cited the accompanying publication – and the many more that did not – you get a good idea what outputs were used for.
When the Cambridge Analytica scandal hit earlier this year, I expected trouble. Although the data doors that allowed CA to gather the data they did were closed in 2015 already, Facebook had to react. A new API version and stricter terms were introduced and, most importantly, all apps would again have to submit to a review. Netvizz was also part of a group of “suspicious applications” that received particular scrutiny concerning their past activities (here is the questionnaire I was asked to fill out). This certainly makes sense, since I could have easily created a treasure trove of data by storing all the data requested through the app.
Looking forward, however, was the new permission system and the connected app review. In order to continue to retrieve data for Pages, the new Page Public Content Access permission is needed. As always, the use cases for getting this permission are rather vague and make no room whatsoever for academic research. Since many researchers and students are relying on Netvizz, I wanted to give the app review a shot.
Such app reviews are not some interactive dialogue with a human operator, but a bunch of form fields and the request to record a screencast that explains how permissions are used. This then normally yields a canned response, not some explanation that actually reflects the submitted app. There is no email address, no appeal process, no “relationship” other than the pre-formatted interface and some pre-formatted answer.
After an initial refusal because “permissions data must be visibly used within your app”, which I think means that you cannot just provide data download, I added a bunch of in-app visualization modules and tried again. After two more attempts, I ended up with a one sentence explanation in the form of a “note from your reviewer”:
While I am not sure what this actually means in practice, it seems clear that the single-page, data-download focused approach Netvizz has been taking is not part of what Facebook wants on its platform in the future. There is maybe room for some widget-driven page comparator and maybe there is a way to sneak out some data that way, but it would go too far to try and build something like this in this structurally opaque and unhelpful environment. So this is where it ends, at least for the moment. Netvizz still functions as I am writing this (try out the new page like visualization module, which allows you to interactively crawl into the network of liked pages, it’s kind of cool) and may very well continue to live a zombie existence in the future – I do not know and have no means to find out.
After tweeting about this process, I not only received a lot of support, but also suggestions to launch some kind of campaign, possibly on the basis of the academic outcomes Netvizz has helped with or enabled. Maybe such an attempt could work, but I personally doubt it since academic research is set to be funneled into new institutional forms that offer more control than API-based data access. Maybe Axel Bruns’ open letter initiative can develop into something like this. I surely hope it does, but for myself it is time to take a break.
Running after too many API changes and jumping through too many bureaucratic hoops over the last three years has made it clear how unsustainable my attempts to provide a tool like Netvizz have really been. Since all my technical work is done in my spare time (I receive no teaching discharge or other compensation from my university) it is simply not possible any more to navigate an environment that has changed so much from the happy go lucky mashup days that started this process. I enjoy programming like few other things in life and even decided to take my university job part time – losing more than half of my salary in the process – to be able to keep doing it, but the increasing hostility toward independent research is creating demands that can only be tackled through more sustainable institutional forms. If any successful resistance to these developments can be mobilized, it will require substantial support from organizations capable of investing resources that may not lead to immediate deliverables or quantifiable publications. At least in my immediate environment, I do not see such forms emerge.
All of this may change again, but for the moment I will take a step back from Netvizz to focus on my more conceptual interests. Facebook will continue to be scrutinized in different forms and through different methods, but the idea that independent research of a 2+ billion user platform just got a lot harder should make us at least a little uneasy.