Social Computing 2

From CS2610 Fall 2016
Jump to: navigation, search

slides

Readings

Reading Critiques

Zhenjiang Fan 16:32:08 11/8/2016

Predicting Tie Strength With Social Media:::::::::::::::::::::: Since people usually use false names online, so I am wondering how the paper select their experiment data? The paper has revealed a specific mechanism by which tie strength manifests itself in social media. Many paths open from here. Social media designers may find traction fusing a tie strength model with a range of social media design elements, including privacy controls and information prioritization. The follow-up interviews suggest profitable lines of future work. It hopes that researchers in this field will find important new theoretical questions in this work, as well as opportunities to use tie strength to make new conclusions about large-scale social phenomena. The finding of the paper is highly doubtable since people usually use false names online and people may give wrong answers when they are interviews about their true identity.::::::::::::::::::::::::::::::::::::::Design Lessons from the Fastest Q&A Site in the West:::::::::::::::::::::::::::::::::To abstract some lessons from Q & A website is quite challenging since you have to know all the information necessary, for example, who asked the question, who answered the question, in what context the question was asked, whether the time spot the question was asked is feasible currently, in what ways we can abstract lessons from it, and in what ways we are going to use these lessons? In this paper, it discussed lessons learned from a popular Q&A site for software developers, Stack Overflow. Analysis of the patterns of user interactions with the site helped to highlight some of the more common user groups and behaviors. A qualitative study helped to explain these behaviors and shed some light on the reasons for the site’s success.

Keren Ye 14:15:56 11/9/2016

Predicting Tie Strength With Social Media The paper present a predictive model that maps social media data to tie strength. The model builds on a dataset of over 2,000 social media ties and performs quite well, distinguishing between strong and weak ties with over 85% accuracy. To assess the tie strength, the paper firstly introduce the dimensions of tie strength. Tie strength has at least seven dimensions and many manifestations, including intensity variables, intimacy variables, duration variable, Reciprocal Services Variables, Structural Variables, Emotional Support Variables, Social Distance Variables. In addition, Demographic and Usage Variables is also mentioned in the paper. Based on these dimensions, the authors build a linear model to predict the tie strength. Then in the following paragraph the authors introduce their experiments and results. The paper answered the question how do users relate to one another in social networks. Yet further challenge is still there. Design Lessons from the Fastest Q&A Site in the West The paper tries to understand the success of Stack Overflow, which is a Q&A system for technical domain. It argues that it is not primarily due to an a priori superior technical design, but also to the high visibility and daily involvement of the design team within the community they serve. At the very beginning, the authors proposed three factors that they believe were critical to the success of Stack Overflow (SO): 1) Making competition productive; 2) Credibility in the community; 3) Evolutionary approach to design. Besides that, the paper states that data analysis and social computing are important factors. In the next part, the paper introduces how Stack Overflow works, and how does it perform. Given the good performance, the authors give qualitative study. Their study shows that SO’s analysis of the patterns of user interactions with the site helped to highlight some of the more common user groups and behaviors. The probable reasons for the success include: 1) founders’ tight involvement with the community, 2) highly responsive and iterative approach to design, and 3) a system of incentives that promoted desirable user behavior.

Steven Faurie 16:59:21 11/9/2016

Steve Faurie: Predicting Tie Strength with Social Media: This paper presents a study that developed a way to predict the strength of ties between users of social media. They describe strong ties and weak ties. Strong ties would consist of relationships with people like family and close friends. Weak ties might be acquaintances or coworkers. They used measurements like links exchanged, number of mutual friends, inbox thread depth etc. to try to predict relationship strength. To evaluate their predictive model they asked users to log into Facebook and rate their tie strength to other individuals. They then took measurements of the predictive variables they were interested in. They compared their model to the ratings the participants told the experimenters. They found some interesting findings, for instance inbox thread depth negatively affects tie strength. Additionally they found educational attainment level was a strong predictor of tie strength. Users with similar levels of education were more likely to be friends. There findings were interesting and I am pretty sure services like Facebook now attempt to evaluate the strengths of these ties to try to make a user’s news feed more interesting…. Design Lessons from the Fastest Q&A Site in the West: This paper looks at Stack Overflow, a well known question and answer site for software developers. The paper wants to examine why the site is so successful. They cited the fast answer times and reputation system the site employs as big contributors to the success of the platform. Additionally they note that the design philosophy of the site’s founders plays a major role in the ongoing success of Stack Overflow. They note the reputation system is sort of like a game for users, the founders were well known and respected and additionally the site developers were responsive to criticism and suggestions from the user base. The paper goes on to describe how successful SO has been. Almost anyone who develops software is aware of the site. And they discuss the community moderation along with the speed at which answers are given. I thought some of the short comings of SO were interesting. The types of questions that are difficult to get answers on are ones that relate to obscure technologies, could be tedious to answer, unreproducible errors, and questions that don’t have a clear answer and could lead to a discussion about preferences etc. When discussing some successful aspects of the site they noted there were answers instead of discussion. I know for a fact that developers prefer to just see an answer than decipher the nuances of different approaches. Especially when they’re just trying to get something done for work. Discussing the reputation system as a game was also an interesting description of how the site works.

nannan wen 18:50:03 11/9/2016

Predicting Tie Strength With Social Media by Gilbert, E. and Karahalios, K. Review: This paper is a literature which shade some light between theory and practice of using the social media to accurately predict the tie strength. In this paper, the author presents a predictive model which can extract interesting effects from the dataset. The author argue that the ability to have a reliable prediction of strong and weak ties could be very useful for social media purposes, so that websites can use this information for the purpose of making the interaction between user and interface more informative. According to the user, the proposed model is accurate, the rate of successful is 85% of the time. Following that, the author discussed discussed the predictive variables that they used throughout user question and Facebook profiles, including detailed descriptions of how these variables were gathered for each other. Then they described the linear model that they used to predict tie strength with their friends. I have been using social media to contact friends a long time ago, if an application can detect the tie that I have with some friends, and then automatically put their posts or status in a priority on my page, it’ll be great. The factors they take into account when calculating ties is amount of time, intimacy, intensity and reciprocal service at first, but then they find that these four factors alone can not represents well about the tie between people. But after some other experiments, they finally defined what the variable should be. Overall, I think the paper proposed a good model in predicting tie strength in the future. But as the author noted, the social network that is being examined is specific in a sense that the members are from same community, so the study might not generalize well to other studies. ---------------------------------------------------------------------Design Lessons from the Fastest Q&A Site in the West by Lena Mamykina et al, Review: In this paper, the author described Stack Overflow, which is a very popular Q&A website for programmers and software engineers to analyzes its design and evolutions. First, the author provided some statistics which shows that Stack Overflow is a successful model. The answer to a question is closer to the time a person were to answer the question. Another proof for the successfulness is the monthly visitor, it greater than 7 million, which is a huge number. They think why it is successful are the following reasons, the first one is “making competition productive”, the second one is “credibility”, the last one is “evolutionary approach to design”. They also list the actions they took to fullfill these requests. After that, they illustrated some analysis methods to analyze this website, they find an interesting fact which is extrinsic motivators. Which is when you add monetary rewards, it can transform user’s sense of system from social interaction space to more formal transaction space. One of the challenges they find is that it took a long time for novice users to get potential status. Overall, this paper is very interesting to me, because I know this website, but not well, after reading this paper, I think I understand this website a little bit more.

Alireza Samadian Zakaria 23:03:08 11/9/2016

Predicting Tie Strength with Social Media: In this paper, the authors have studied relationships in social media and in reality; then they have tried to predict tie strength by using data from social media. At first, the paper talks about one of the definitions of ties and differences between strong tie and weak tie. Granovetter has proposed four dimensions for tie strength: the amount of time, intimacy, intensity and reciprocal services. However, the paper proposes other dimensions for strong ties in practice: communication reciprocity, processing at least one mutual friend, recency of communication, and interaction frequency. They designed an experiment for prediction and in this experiment, they had 7 groups of independent variables which were Intensity, Intimacy, duration, reciprocal services, structural, emotional support, and social distance. Then they asked 5 questions from their participants regarding their friends to measure tie strength. At the end, they found the most influential group of independent variables was intimacy and the most important variable was days since last communication. At the end, the author discusses some of the statistical factors in their results. ----------- The second paper is about one of the most successful question and answer website, Stack Overflow, and the reasons for its success. According to this paper, 92.6 percent of the questions are answered in this website (half of them in less than 11 minutes); this with some other factors such as the quality of the answers and community guidelines are the main reasons of its effective functioning. Another important feature of this website is that their designers also use this system so they are actually somehow its user too. This leads to some factors that are considered to be critical to the success of SO such as productivity of competitions, credibility in the community, and evolutionary approach to design. One of the interesting points about this website is that it has been fast since the start of its activity and the median of the first answer time has not changed since then. However, some questions are answered slowly or remain unanswered; according to the hypothesis of the authors, there are some possible reasons for that such as obscurity of their technology or being tedious to answer.

Xiaozhong Zhang 23:52:05 11/9/2016

Predicting Tie Strength With Social Media In this paper, the author present a predictive model that maps social media data to tie strength. The model builds on a dataset of over 2,000 social media ties and performs quite well, distinguishing between strong and weak ties with over 85% accuracy. The author complement these quantitative findings with interviews that unpack the relationships the author could not predict. The paper concludes by illustrating how modeling tie strength can improve social media design elements, including privacy controls, message routing, friend introductions and information prioritization. The author hope that researchers in this field will find important new theoretical questions in this work, as well as opportunities to use tie strength to make new conclusions about large-scale social phenomena. Some limitations of the paper are: The specific predictive variable coefficients may not move beyond Facebook. That being said, this work looks only at one social media site, at one time, using data available through the browser and lacks evaluating the utility of "behind-the-scenes" data and to work contrasting these findings with other social media. Design Lessons from the Fastest Q&A Site in the West This paper analyzes a Question & Answer site for programmers, Stack Overflow, that dramatically improves on the utility and performance of Q&A systems for technical domains. Over 92% of Stack Overflow questions about expert topics are answered -- in a median time of 11 minutes. Using a mixed methods approach that combines statistical data analysis with user interviews, the author seek to understand this success. The author argued that it is not primarily due to an a priori superior technical design, but also to the high visibility and daily involvement of the design team within the community they serve. The author further stated that this model of continued community leadership presents challenges to both Computer-supported cooperative work (CSCW) systems research as well as to attempts to apply the Stack Overflow model to other specialized knowledge domains. The paper conducted analysis of the patterns of user interactions with the site and found some of the common user groups and behaviors. Then through a qualitative study, the paper found some reasons to explain the behaviors and the site's success. These reasons include 1) founders' tight involvement with the community, 2) highly responsive and iterative approach to design, and 3) a system of incentives that promoted desirable user behavior.

Debarun Das 7:36:48 11/10/2016

“Predicting Tie Strength With Social Media”: With the rapid use of social media in the current world, this paper serves as an interesting read. This paper aims to determine the social media tie strength between two friends in social media. The purpose of such determination is that it will help in bettering many aspects like information prioritization, privacy control etc. They discuss about the seven standard dimensions of tie strength. This method used the data from Facebook to recruit participants to answer research questions, that will help determine tie strength. It then goes on to describe the different variables that can be defined here. Finally, they discuss about the results, which are quite favorable. This being one of the earlier papers that discussed this, paved the way the way for further research in this field. --------------------------------------- “Design Lessons from the Fastest Q&A Site in the West”: This paper mainly studies about the Stack Overflow Q&A system in which about 92% of expert topic questions are answered at a median of almost eleven minutes. This paper aims to understand this success using a mixed methods approach. This means that they combined statistical data analysis with user interviews. As a conclusion, this paper determined the three reasons for the success of Stack Overflow: i) Founder’s tight involvement with community, ii) highly responsive and iterative approach to design, iii) a system of incentives for desired user behavior. One of the main thing it tries to understand is the importance of socio-cultural context in social computing platforms.

Anuradha Kulkarni 7:44:48 11/10/2016

Predicting Tie Strength With Social Media: It has been observed that the Social Media treats all the user at the same aspect be it a trusted friend or a stranger. In reality, it is different where relationships fall everywhere along this spectrum. This paper bridges this gap between the theory aspect and practical aspect. This paper presents a model to quantify social tie strength between friends on social media, specified in Facebook. I agree with the author that Tie strength can help many ways in social network, like privacy controls, friend introductions and information prioritization. However, this paper fails to address how to use this information in designing a new social network. The model builds on a dataset of over 2,000 social media ties and performs quite well, distinguishing between strong and weak ties with over 85% accuracy. The way authors use script to collect data is not enough since the script was run in client side. Overall this paper was a good read and informative. It started off explaining the concepts and variables very well. It also showed some statistics analysis. It gave a good perspective of social media and its future implications. --------------------------------------------------------------------------------------------------------------------------------------------------- Design Lessons from the Fastest Q&A Site in the West: The paper mainly discusses about Stack Over Flow website and reasons why it's so successful. Stack overflow is actively used amongst programmers and it helps in giving quick answers to the problem that we are stuck at. Stack Overflow is a Question & Answer site for programmers. This papers shows that this improves the utility and performance of Q&A systems for technical domains. At a very higher rate the median time taken to answer expert topics is observed to be 11 min, which in my opinion is a quick response. There is no doubt that with such quick response time, the number of visits to this site will be greater. Stack overflow has short thread length, however it's very concise and useful, especially when they allow comments on each solution. These help users a lot in order to find the correct answer for their own problems. The paper presents the three reasons they believe were instrumental for the success of Stack Overflow. First one is “making competition productive”, which was done by making some game mechanics that led the users to be active. Second one is “credibility”, which was achieved by special status and badges to the dedicated users. The last one is “Evolutionary approach to design” which was achieved by the prompt reaction of the design team for the feedback they get through the meta site. I am in par with all these three reasons. The paper is informative, interesting and has been supported with lots of graphs.