sequence matching

on social media



{HHR, 05-Feb-2018}

I’m thinking about the meaning of attention also with respect to our work with the algorithm. We have of course a moving focus, when we write code. We perfectly know that we have very limited cognitive capabilities to understand the system we're building in their entirety. I'm talking about even the small systems that we create: they easily go beyond the point where we can have a totality of focus, it’s simply not possible. I’m trying to understand how that could be something that it’s very relevant to what we’re doing: the idea that there must be rhythm or motion in terms of the focus. This may refer to the question regarding whether we are purely sequential or if there’s some second layer of parallelism, the idea to recurrently return to something from before.

{origin: conversation transcript, keywords: [attention, focus, algorithm, cognition]}

{RK, 29-Aug-2017}

Do you know anyone who does any work on automatic searches of social media? I am interested in identifying search keys that get "slow but steady" use (perhaps on youtube).  Most of focus of attention in relation to social media goes to "viral" elements.  I am curious about what searches persist on the margin of the social media economy of attention.

Within the broader theme of the project, most social media is structured around the idea that a user's desires are reflected in their past consumption.  So, part of the idea here is to create meta-searches that cannot be expressed in those terms

{origin: email, keywords: [social media, attention, search]}


{hhr, 30-Aug-2017}

i'm not aware of anyone in our lab having done work with automated searches on social media. i remember there was a sonification competition with respect to twitter, you know when there was this hype about the twitter API. i have very limited experience myself, i know the basics of REST queries, having written a query library, and i guess most of the social networks that offer APIs are kind of similar in the general approach. in terms of trends, i think there might be a few things, i remember google trends: - perhaps they have some sort of API (this question suggests that 'no' but there are workarounds: but again, i think these services will mostly focus on the 'viral' aspects and not the margins of what is happening. the problem is that most social media are either proprietary and closed or open source but decentralised (and thus probably without such query facilities). i would imagine that if you already have a particular query term, then you can perhaps trace how it evolves over time, using temporal constraints in the search for example. i would also think that it's much easier to get hold of data that people publish instead of their searches, as that
data by definition is publicly available. so you could probably crawl twitter feeds and make statistics across time bands, and so forth.

{origin: email, keywords: [social media, search, trends]}

meta: true
persons: [RK, HHR]
kind: fragment