Project »Twitter

Difference between r1.239 and the current

@@ -1,46 +1,47 @@
#keywords Research on social networks and message diffusion on Twitter
#title Research on social networks and message diffusion on Twitter

***
* The below was the message sent to Twitter API team (refer to the message thread [http://groups.google.com/group/twitter-development-talk/browse_thread/thread/36a11ab23b42f22/a14e0f1a5691cbc3?lnk=gst&q=hkimscil#a14e0f1a5691cbc3 #1]). The request for the whitelist, which is essential for the data collection was denied (the message thread [http://groups.google.com/group/twitter-development-talk/browse_thread/thread/b63cbd68138b7a5e/7c860bee68479bfa?lnk=gst&q=hkimscil#7c860bee68479bfa #2].
* The point that they made is that the plan should be some sort of application building - not social scientific studying. . . .
__Research questions__
1. What does affects agenda in Twitter? What are the factors that influence the agenda in tweets? How are they compared to agenda in traditional mass media?
1. It is very interesting and important topic to explore how the message agenda are set in Twitter. Agenda may set differently from those of traditional mass media (newspapers and broadcasting). Or agenda may be similar to those from mass media; but, the diffusion processes may differ from each other. Future journalism activities; corporation public relations and marketing, and so on will be benefited from the study. In order to figure out these, we need to explore the following questions.
1. How are the messages diffused; and what are the structural features of such diffusion?
1. 2 different approaches in message diffusion study
a. based on individual social networks : This approach is __potential__ message diffusion [[FootNote(Previous studies are found at http://an.kaist.ac.kr/traces/WWW2010.html and http://twitter.mpi-sws.org/)]].
a. based on message diffusion pattern: This approach is __actual__ message diffusion.
1. We use '''the second''' approach (b); and try to articulate how the diffusion patterns (if any) can be generalized.
a. refer to the [Twitter_0.8] document written in Korean.
a. In sum, the study aims at categorizing some distinguishable message-distribution lines, and seek out what their characteristics are.
1. We also plan to identify structural features of social networks that are actually involved in message diffusion processes.
1. 무엇을
* 무엇을 (리)트윗 하는가?
* 어떤 종류의 메시지를 (리)트윗하는가?
* 소식 전달 중심인가?
* 개인 관계망 관리 중심인가?
* 사회적인 가치를 갖는 메시지가 우선시되는가 혹은 개인사적인 메시지가 더 많이 유통되는가?
* 이는 __왜 하는가와도 연관__된 질문 -- 같이 묶어서 서베이 문항을 발전 시킬것
* 누구와 하는가와도 약하게 연결 . . . .
1. 누구와
* ego network 관점에서 누구의 트윗을 리트윗 하는가?
* ego network 관점에서 관계는 어떻게 세분화 될까?( 친구, 동료, 관심분야 단체(연구소, NGO, 종교...), 뉴스매체, 뉴스매체 종사자, 관심분야 전문가 및 종사자, 상업회사, 유명인) - 카테로리별 분류 . . . .
* __어떤 종류의 관계가 뚜렷하게 나타나며, 이런 관계는 어떻게 인식되고 있는가? (약한 연결의 강함과 같은 의미에서 . . . . )__
=> ego network는 다양한 관계를 갖고 있을 것이며, 특정 사안에 특정 네트워크가 반응하는 마치 뉴런과 같을 것이라 예상
1. 왜 하는가?
* 왜 하는가? uses and gratifications + indepth study (리트윗 하는 글의 속성, 그 글을 리트윗하는 동기, 최종적으로 추구하는 가치)
- 마이크로블로그에서 메시지 전달자의 가치체계는 어떻게 구성되는가?
- 메시지 전달시 반영되는 속성은 무엇인가?
- 메시지 전달자가 각각의 속성을 통해 얻고자 하는 결과는 무엇인가?
- 메시지 전달을 통해 궁극적으로 추구하는 가치는 무엇인가?

__Data__
1. We '''collect trending topics''' via Local trends (in Search API): We need daily or weekly trends; but they are not available. Hence, we may track the trending topics every 3 hours in a day (8 total - they become the trending topic for that specific day) or so. We will end up about 20 different topics a day. We use these topics for tracking message data (tweets).
1. With trending topics, we '''collect related messages for two weeks''' (via Search API or Streaming API [[FootNote(We are not sure which one would be better, yet.)]]). For example, we may have 210 trending topics (30 per day * 7 days). Each topic may have 50,000-many messages (since they are collected from local trends, the number of related messages will be significant).
1. We '''identify the owners''' of the collected messages; and via REST API, we collect information about users who participated in the trending topics and their social graph data. In doing so, we use various methods of REST API.
* programming?
필요한 것들 . . . . . 정리할 것. . . . .
* survey?
* programming + survey ?

1. We also try to collect headlines information of major local newspapers (such as NY Times (for NY local)). This data set is to be compared with the local trending topics in Twitters.
__Reading__
* http://bklove.info/1116
* [Twitter/References]
* [TwitterResearch]
* [TwitterResearch/ResultSage01]

__Scripts__
See [Twitter_Pattern Twitter Pattern] for data collection scheme.
See [Twitter Libraries]

__Requests__
1. We have set up about 22 computers for collecting data sets. The IP addresses range from xxx.xxx.xxx.xxx to xxx.xxx.xxx.xxx. 22 computers are used in order to make data gathering process fast and easy.
1. We need to be '''whitelisted''' in SEARCH API: We need to collect all the messages (Tweets) of the trending topics that we choose to use.
a. 210 topics (or less) = About 30 topics a day for over two weeks
a. each topic may have many messages = 50,000 - whatever number.
a. We will use scripts to collect local trend topics; then use Search API to collect related messages.
1. We need to be '''whitelisted''' in REST API: We need to collect public information about users who have participated in message diffusion.
a. We identify the owner of the messages gathered in the previous step.
a. Collect public information about them and their social network graph data (and some others).

__Contact Information__
related ID: tweet_study

__Contact Information__
Hyo Kim, Associate Professor, Department of Digital Media Ajou University
- [mailto:hkim@commres.org hkim at commres dot org] or [mailto:hkimscil@ajou.ac.kr hkim at university address]
- Referred research works (in English):


Research questions
  1. 무엇을
    • 무엇을 (리)트윗 하는가?
    • 어떤 종류의 메시지를 (리)트윗하는가?
      • 소식 전달 중심인가?
      • 개인 관계망 관리 중심인가?
      • 사회적인 가치를 갖는 메시지가 우선시되는가 혹은 개인사적인 메시지가 더 많이 유통되는가?
    • 이는 왜 하는가와도 연관된 질문 -- 같이 묶어서 서베이 문항을 발전 시킬것
    • 누구와 하는가와도 약하게 연결 . . . .
  2. 누구와
    • ego network 관점에서 누구의 트윗을 리트윗 하는가?
    • ego network 관점에서 관계는 어떻게 세분화 될까?( 친구, 동료, 관심분야 단체(연구소, NGO, 종교...), 뉴스매체, 뉴스매체 종사자, 관심분야 전문가 및 종사자, 상업회사, 유명인) - 카테로리별 분류 . . . .
    • 어떤 종류의 관계가 뚜렷하게 나타나며, 이런 관계는 어떻게 인식되고 있는가? (약한 연결의 강함과 같은 의미에서 . . . . )
      => ego network는 다양한 관계를 갖고 있을 것이며, 특정 사안에 특정 네트워크가 반응하는 마치 뉴런과 같을 것이라 예상
  3. 왜 하는가?
    • 왜 하는가? uses and gratifications + indepth study (리트윗 하는 글의 속성, 그 글을 리트윗하는 동기, 최종적으로 추구하는 가치)
      - 마이크로블로그에서 메시지 전달자의 가치체계는 어떻게 구성되는가?
      - 메시지 전달시 반영되는 속성은 무엇인가?
      - 메시지 전달자가 각각의 속성을 통해 얻고자 하는 결과는 무엇인가?
      - 메시지 전달을 통해 궁극적으로 추구하는 가치는 무엇인가?

Data
  • programming?
    필요한 것들 . . . . . 정리할 것. . . . .
  • survey?
  • programming + survey ?

Reading
Scripts
See TwitterLibraries

Requests


Contact Information

Hyo Kim, Associate Professor, Department of Digital Media Ajou University
- Mhkim at commres dot org or M hkim at university address
- Referred research works (in English):
[http]Comparing Academic Hyperlink Structures with Journal Publishing in Korea
[http]Configurations of Relationships in Different Media

Hyon Seok Shin, Research Assistance, Department of Digital Media Ajou University
Mhshin at commres dot org
Valid XHTML 1.0! Valid CSS! powered by MoniWiki
last modified 2012-05-08 14:46:26
Processing time 0.0425 sec