[Bigbang-dev] Bipartite graphs of working group and affiliations

[Corinne Cath]

Hi Seb,

Thanks! This is very exciting work. One thing I was wondering about is
legibility for non-academic audiences.

These kind of network graphs are only legible to a very small part of the
population for whom it might be interesting. Maybe this is a bit premature,
but I would like to raise that we might need to do quite a bit of
translational work to make these graphs useful for advocacy purposes.

I don't necessarily have much experience in doing that, but happy to
suggest people that do and keep the convo doing!

Kind regards,

Corinne

On Tue, Jun 9, 2020 at 7:19 PM Sebastian Benthall <sbenthall at gmail.com>
wrote:

> Thanks for this, Nick.
>
> That reminds me: I'm on the hook to get Nels a list of top 100
> affiliations for categorization.
>
> This visualization is noting affiliations of draft authors only. Not
> general discussion participants. I'm not sure if the IETF data tracker has
> lighter weight documents in it that might pick up wider participation. Any
> ideas what those would be called? Are "comments" a tracked document type?
>
> One thing that *might* pop out of a more extensive mapping (longer time
> scale, etc) is some kind of modular structure to the network. It might
> *not* be there, an it could be one big hairball that doesn't tell us much.
> But I wonder if you would expect, based on qualitative experience and
> intuition, any clustering?
>
>
> On Tue, Jun 9, 2020, 1:11 PM Nick Doty <npdoty at ischool.berkeley.edu>
> wrote:
>
>> Awesome!
>>
>> My initial thought was just that it seems so sparse. httpbis is a massive
>> effort (based only on my dropping in to WG meetings occasionally, I haven’t
>> been following the list), how can it only have 2 organizations involved?
>> Maybe that’s a side effect of your using a short date range?
>>
>> I also noticed the entity resolution issue (with Cisco).
>>
>> One issue with connection working groups to companies is that it might
>> become overly dense. When we did this just with individuals, there was
>> already quite a bit of overlap/connection, with only occasional groups that
>> were pushed to the outside of the diagram because many individuals on those
>> lists only contributed to them. But with looking at entire companies, it
>> seems more likely that you’ll see the large vendors connected to a wide
>> range of WGs across multiple areas and so I’m not sure what the
>> visualization will be able to show you.
>>
>> Thanks for sharing this work in progress. I think affiliation can be such
>> a promising area, particularly if we can start to classify the affiliations
>> (corporate, non-profit, academic, etc.).
>>
>> Cheers,
>> Nick
>>
>> > On Jun 8, 2020, at 3:10 PM, Sebastian Benthall <sbenthall at gmail.com>
>> wrote:
>> >
>> > Hello,
>> >
>> > I've been working with the IETF data tracker to produce a visualization
>> of affiliations and working groups. I'd like to run it by you all for
>> design feedback and ideas moving forward.
>> >
>> > This is a bipartite graph of affiliations and working groups for the
>> period between 2010-01-01 and 2010-04-01:
>> >
>> > <image.png>
>> > Aside from entity resolution issues (Juniper/Juniper Networks), I
>> wonder if anything pops out at you or of the visualization might contain
>> more information in a useful way.
>> >
>> > Cheers,
>> > Seb
>> > _______________________________________________
>> > Bigbang-dev mailing list
>> > Bigbang-dev at data-activism.net
>> > https://lists.ghserv.net/mailman/listinfo/bigbang-dev
>>
>> _______________________________________________
> Bigbang-dev mailing list
> Bigbang-dev at data-activism.net
> https://lists.ghserv.net/mailman/listinfo/bigbang-dev
>


-- 
Corinne Cath - Speth
Ph.D. Candidate, Oxford Internet Institute & Alan Turing Institute

Web: www.oii.ox.ac.uk/people/corinne-cath
Email: ccath at turing.ac.uk & corinnecath at gmail.com
Twitter: @C_CS
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ghserv.net/pipermail/bigbang-dev/attachments/20200611/2006ab3a/attachment.html>