15 ms·

It's unparalleled in that there is very little that's even similar out there. I'm less convinced of its actual value. The Twitter-based papers I've read seem t

It's unparalleled in that there is very little that's even similar out there. I'm less convinced of its actual value. The Twitter-based papers I've read seem to all have two things in common: First, they all use Twitter data. Second, they all seem to desperately want to present observations about how people interact on Twitter as being universal, even though nobody seems to be able to come up with a theoretical or empirical reason why that should be the case that's more compelling than, "We we have to pretend it is in order to get published."

lovamova 10y ago
Similar like in Sublevel?
tootie 10y ago
There's thousands of sites serving the same or similar purpose. Twitter is only unique because it has a lot of users.
- bunderbunder 10y ago
  It's not really about whether or not the site works similarly to Twitter. It's about the neat stuff people can build with something like Twitter data. I don't think you could do something like http://blog.idl.ssrc.msstate.edu/?p=42 http://blog.idl.ssrc.msstate.edu/?p=42 with Plurk data and get anywhere near the same kind of attention for it. For that kind of thing, having lots of users is the key feature of interest.
  - Godel_unicode 10y ago
    Is attention the goal? I've been struck by the number of articles and research projects that are some version of "we got lots of tweets and graphed them" and have totally failed to go the "so what" step. Why should we care about this project? If you want to use this to find power outages using Twitter, why not just do that directly with e.g. grep (also, more importantly, are these outages about which the utility company doesn't already know?). Having big data (related: 12 Million lines of text isn't big) isn't cool in and of itself. Doing things which help with that data would be cool.
nostromo 10y ago
Yes. It also bothers me that quality news sources, like NPR, have fallen into the trap of presenting Twitter as representative of the general population. They often report about how some group of people on Twitter are angry about something -- as if that matters. Reddit has almost as many monthly active users lately, yet no journalist makes the mistake of reporting on Reddit comments as if it's generalizable to the larger population.
- SolaceQuantum 10y ago
  Most of the news reporting I've seen cast on individual Reddit communities. I don't think such a thing can be done as easily as twitter except through hashtags, thus the coverage on twitter being more twitter-wide hashtagging rather than sub-tweet communities.
  - csallen 10y ago
    What's interesting is that, even though they're not well-defined, there certainly are specific communities on Twitter. There's tech, and sports, and science, psychology, politics, music, etc, and each of those can be further divided. I'm sure that if Twitter put in the work, they could suss out all these communities based on who is following who and what they tweet about. (In fact, I'd be surprised if they aren't already doing that.) They could then make these communities accessible to the public. Twitter currently has lists, but the problem is that these lists aren't crowdsourced (to my knowledge). They're maintained by individuals, which is unsustainable, because one person can only do so much work keeping a list accurate and up-to-date.
    - ghaff 10y ago
      I'd be pretty sure that "Who to Follow" is based on some sort of social graph though probably in a simplistic way, e.g. lots of people you follow follow this person and you don't.
      - csallen 10y ago
        Yeah, that's the only thing I've seen from them in this domain. It's just woefully lacking imo. I don't want to pick people to follow individually... I want to pick entire communities of people to follow.
eveningcoffee 10y ago
I heard one of the "big data social scientists" to claim that they can now replace the population with the sample (as this is what the Twitter really is) because of, you know, big data.
- dapz 10y ago
  If you have a demographic model of twitter users and appropriately re-weight, why is this a bad approach to take?
  - eveningcoffee 10y ago
    If I understood correctly then this was considered as a not necessary step. Now given the Twitters nature then is there such a model that is even approximately accurate? Or in other words, what relevant data Twitters collects and how accurate it is?
  - Godel_unicode 10y ago
    Because you don't (and probably can't) know that this model is accurate? Also, there are a lot of populations which are not present on Twitter at all, and more which are not present in meaningful numbers.
    - dapz 10y ago
      Such models presumably can be tested in some sense. E.g. build some sort of predictive model of elections in each state using Twitter users as a surrogate, see how well you do. I'm not claiming that such a model is guaranteed to be perfect or even as good as a pollster, but it sounds in principle like a reasonable thing to do, if you take appropriate precautions to reduce bias.
      - eveningcoffee 10y ago
        At least one can try definitely.

lovamova 10y ago

Similar like in Sublevel?

tootie 10y ago

There's thousands of sites serving the same or similar purpose. Twitter is only unique because it has a lot of users.

bunderbunder 10y ago
It's not really about whether or not the site works similarly to Twitter. It's about the neat stuff people can build with something like Twitter data. I don't think you could do something like http://blog.idl.ssrc.msstate.edu/?p=42 http://blog.idl.ssrc.msstate.edu/?p=42 with Plurk data and get anywhere near the same kind of attention for it. For that kind of thing, having lots of users is the key feature of interest.
- Godel_unicode 10y ago
  Is attention the goal? I've been struck by the number of articles and research projects that are some version of "we got lots of tweets and graphed them" and have totally failed to go the "so what" step. Why should we care about this project? If you want to use this to find power outages using Twitter, why not just do that directly with e.g. grep (also, more importantly, are these outages about which the utility company doesn't already know?). Having big data (related: 12 Million lines of text isn't big) isn't cool in and of itself. Doing things which help with that data would be cool.

nostromo 10y ago

Yes. It also bothers me that quality news sources, like NPR, have fallen into the trap of presenting Twitter as representative of the general population. They often report about how some group of people on Twitter are angry about something -- as if that matters. Reddit has almost as many monthly active users lately, yet no journalist makes the mistake of reporting on Reddit comments as if it's generalizable to the larger population.

SolaceQuantum 10y ago
Most of the news reporting I've seen cast on individual Reddit communities. I don't think such a thing can be done as easily as twitter except through hashtags, thus the coverage on twitter being more twitter-wide hashtagging rather than sub-tweet communities.
- csallen 10y ago
  What's interesting is that, even though they're not well-defined, there certainly are specific communities on Twitter. There's tech, and sports, and science, psychology, politics, music, etc, and each of those can be further divided. I'm sure that if Twitter put in the work, they could suss out all these communities based on who is following who and what they tweet about. (In fact, I'd be surprised if they aren't already doing that.) They could then make these communities accessible to the public. Twitter currently has lists, but the problem is that these lists aren't crowdsourced (to my knowledge). They're maintained by individuals, which is unsustainable, because one person can only do so much work keeping a list accurate and up-to-date.
  - ghaff 10y ago
    I'd be pretty sure that "Who to Follow" is based on some sort of social graph though probably in a simplistic way, e.g. lots of people you follow follow this person and you don't.
    - csallen 10y ago
      Yeah, that's the only thing I've seen from them in this domain. It's just woefully lacking imo. I don't want to pick people to follow individually... I want to pick entire communities of people to follow.

eveningcoffee 10y ago

I heard one of the "big data social scientists" to claim that they can now replace the population with the sample (as this is what the Twitter really is) because of, you know, big data.

dapz 10y ago
If you have a demographic model of twitter users and appropriately re-weight, why is this a bad approach to take?
- eveningcoffee 10y ago
  If I understood correctly then this was considered as a not necessary step. Now given the Twitters nature then is there such a model that is even approximately accurate? Or in other words, what relevant data Twitters collects and how accurate it is?
- Godel_unicode 10y ago
  Because you don't (and probably can't) know that this model is accurate? Also, there are a lot of populations which are not present on Twitter at all, and more which are not present in meaningful numbers.
  - dapz 10y ago
    Such models presumably can be tested in some sense. E.g. build some sort of predictive model of elections in each state using Twitter users as a surrogate, see how well you do. I'm not claiming that such a model is guaranteed to be perfect or even as good as a pollster, but it sounds in principle like a reasonable thing to do, if you take appropriate precautions to reduce bias.
    - eveningcoffee 10y ago
      At least one can try definitely.