[Air-L] Lost youtube comments

Bernhard Rieder berno.rieder at gmail.com
Wed Jan 17 07:22:58 PST 2018

Dear Muira,

I would like to push back a little against something you are saying in your message, namely that Webrecorder (or any other tool for that matter, including the YouTube Data Tools mentioned below, which I develop and maintain) allows you to create a "reliable archive". I think that this is a problematic starting point, in particular for data retrieved from corporately owned platforms via data interfaces designed for purposes other than research goals. The YouTube API, for example is highly reactive to privacy settings, allows admins to delete and moderate comments, and does not always provide all comments for thread chains for a variety of reasons. I would argue that the "reliability" of data needs to be established for each research project sparately and cannot be delegated to a tool. This validation (which includes an understanding of limitations) can be done, for example, through comparison between data and interface, by using different API credentials for retrieval, and by supporting findings through two or more approaches.

More specifically, for any data retrieved from corporate platforms, research should - in my view - include robust data critique. My colleagues and I have tried to sketch such a process for Facebook API data here: http://journals.sagepub.com/doi/abs/10.1177/2053951715614980

best wishes,

Bernhard Rieder | Associate Professor | New Media and Digital Culture  
University of Amsterdam | Turfdraagsterpad 9 | 1012 XT Amsterdam | The Netherlands  
http://thepoliticsofsystems.net | http://labs.polsys.net | https://www.digitalmethods.net | @RiederB

> On 17 Jan 2018, at 14:52, Muira McCammon <muira.n.mccammon at gmail.com> wrote:
> Hi, Philippa!
> This is a frustrating issue, to be sure.
> Though I am not very well-acquainted with YouTube's framework/structure and
> know nothing of the Digital Methods Initiative, this is what I can offer in
> terms of advice for future research: Webrecorder.io, an open-source project
> by Rhizome at the New Museum, gives users 5GB of space, which you can use
> to create your own high fidelity web archives.  To give you a sense of how
> this works, last evening, to test things out, I went to YouTube and made a web
> archive
> <https://webrecorder.io/muira/youtubetest/20180117032525/https://www.youtube.com/watch?v=GwN-WixHkAw>
> of this video <https://www.youtube.com/watch?v=GwN-WixHkAw> of NPR's Tiny
> Desk Concert featuring Dessa; as you can see, I included the comments in
> the archive I created.
> What Webrecorder would let you do is create a reliable archive of the
> YouTube comments you want to incorporate into your research. That archive
> would then be accessible, and you could link to it, if need be, long after
> the comments disappeared and/or were deleted. Normally, if you were just
> interested in archiving static material (that is, the comments and not say,
> video content), I would suggest you use Perma.cc, an archiving tool created
> by the Harvard Law Library Innovation Lab. The problem there is two-fold:
> (#1) you'd only get 10 links a month, which isn't much, and (#2) based on
> my own use, it just doesn't deal well with any YouTube content, including,
> the comments. I should say here that there are other web archiving tools
> out there, but Webrecorder.io and Perma.cc are what I know best. Others on
> this listserv may have other/better suggestions!
> If you go down this road, some folks on this listserv might urge you to
> grapple with some of the ethical issues that come with creating web
> archives of YouTube user comments, esp. those that users later delete
> deliberately; maybe someone will weigh in on that matter. Indeed, perhaps
> someone on this listserv has written about the ethics of web archiving
> YouTube comments!
> That being said, at a technical level, if you're trying to keep track of
> thousands of comments or hundreds of YouTube pages, you will probably hit
> the 5GB data limit on Webrecorder sooner than later.  Good news is that
> Rhizome <http://rhizome.org/> just received a grant
> <https://news.artnet.com/art-world/rhizome-1-million-mellon-foundation-grant-1197608>
> of $1 million, so perhaps they'll bump up the space they give new users.
> (In full disclosure, I should say that I received a microgrant from Rhizome
> back in 2017, so I am, in more ways than one, a fan of what they do. Happy
> to help anyone here figure out how to use their tool.)
> Lastly, I'd be curious to know if anyone here, apart from myself, is
> actively using Webrecorder in their research, especially if it's
> qualitative and engaging with deletion on social media platforms. Would
> love to talk to you, if you exist!
> All my best,
> Muira McCammon
> Muira McCammon
> Ph.D. Student <https://www.asc.upenn.edu/people/students/muira-mccammon> /
> Journalist <http://muira.me/>
> Signal: Available upon request
> Annenberg School for Communication
> University of Pennsylvania
> On Tue, Jan 16, 2018 at 5:22 PM, Philippa Smith <philippa.smith at aut.ac.nz>
> wrote:
>> Hi Everyone
>> My latest research project involves linguistic analysis of
>> comments/replies under a specific youtube video.  Interestingly I cut and
>> pasted the comments from a particular youtube url in July last year and
>> saved them in a word document. However in going back to update this in 2018
>> I find that those original comments are no longer accessible in that the
>> site does not automatically load more comments beyond the last three
>> months.  This leave me with an incomplete data set.  I find this situation
>> rather intriguing.   If anyone who is better informed than I am  might
>> enlighten me as to whether this just reflects a change in youtube policy or
>> might relate to settings of the youtube poster - that would be helpful. And
>> if there is any solution - even better!
>> Kind regards
>> Philippa
>> [AUT]<https://www.aut.ac.nz/>
>> Philippa Smith
>> Senior Lecturer
>> Auckland University of Technology
>> [QS rankings]<https://www.aut.ac.nz/qs>
>> [Facebook]<https://www.aut.ac.nz/facebook>[Twitter]<https://
>> www.aut.ac.nz/twitter>[YouTube]<https://www.aut.ac.nz/youtube>
>> P 09 921 9999 ext 8276    E philippa.smith at aut.ac.nz<mailto:
>> philippa.smith at aut.ac.nz>    W aut.ac.nz<https://www.aut.ac.nz/>
>> [https://www.aut.ac.nz/__data/assets/image/0011/571817/teal-strip.gif]
>> _______________________________________________
>> The Air-L at listserv.aoir.org mailing list
>> is provided by the Association of Internet Researchers http://aoir.org
>> Subscribe, change options or unsubscribe at: http://listserv.aoir.org/
>> listinfo.cgi/air-l-aoir.org
>> Join the Association of Internet Researchers:
>> http://www.aoir.org/
> -- 
> *Muira McCammon*
> *@muira_mccammon <https://twitter.com/muira_mccammon>*
> Recently out: "Persisting in Dark Times: Lessons from a War Crimes
> Researcher
> <http://howwegettonext.com/persisting-in-dark-times-lessons-from-a-war-crimes-researcher-8b3504f4b169>"
> in *How We Get to Next *
> _______________________________________________
> The Air-L at listserv.aoir.org mailing list
> is provided by the Association of Internet Researchers http://aoir.org
> Subscribe, change options or unsubscribe at: http://listserv.aoir.org/listinfo.cgi/air-l-aoir.org
> Join the Association of Internet Researchers:
> http://www.aoir.org/

More information about the Air-L mailing list