Social networks and group formation
Note: Below is the translation of the academic research “Social Networks And Group Formation” , which examines some aspects of the selection of groups and the formation of connections in social networks.
Users suffer from an excess of information: on any issue, much more information is now available than a person is able to perceive. As a result, people rely on the knowledge of other people. Among the typical questions “how?”, “What?” And “why?”, The key question is “who?” (And, accordingly, the answer to it). This entails the need to meet and maintain contact with people who can help in each case.
')
In the context of the stated problem, it is also important to understand the education, development and use of social networks on the Internet. A social network is a “group of people (or organizations, or other social units) whose connections are social, namely friendship, collaboration, or the exchange of information” (Garton et al., 1997). Although the Internet is the main source of redundant information, it also allows the user to effectively manage their social networks and thus receive the information that they need.
This area of knowledge is located at the intersection of such sciences as computer science, sociology and mathematics. Its researchers are studying the use of social networks for public and business purposes using information and communication technology (ICTs) , including the Internet. This article discusses how social networks, by taking advantage of such technologies — especially Internet technologies — are created, developed, and profitable.
Social networks on the Internet are an extensive field for research, and any scientific review will in one way or another be based on a set of academic publications. This article focuses on the latest research on network center ( centrality ), the strength of connections, individuality, trust, activity and benefits. Undoubtedly, the article is somewhat difficult to understand, but it is intended to convey to developers several useful ideas on the implementation of web applications that actively use social networks.
For the first time social networks were investigated in the late 40s of the last century. With the advent of the Internet, online communities and relevant websites, their importance has only increased. Any self-respecting author who claims to be objective, in his article will have to mention the sociologist Mark Granovetter and the mathematician Linton C. Freeman , who wrote the basic materials even before the widespread use of the Internet.
Granovetter (1973) stated that within social networks, weak links are much more important than strong ones. He explained this by the fact that through weak links information “seeps” faster. This, in his opinion, is due to the fact that weak links are “more important for individual users when they are“ infused ”and interact in the community, whereas as a result of strong links a close local group is formed.
In his doctoral thesis, Granovetter showed that people are more likely to find work because of their weak ties, rather than strong ones. Weak ties are links with people whom we hardly know, with whom we don’t share our feelings and do not maintain relationships; but they turn out to be the most useful. This is due to the fact that people with strong connections share mostly the same data or resources, thus they are less useful to each other.
Along with them, Granovetter identified the missing connections (also called “ nodding ”) - such connections lack the emotional component, time, trust and reciprocity - as weak. Someone living with you on the same street, to whom you nod every day, will be just an example of such a connection. Missing communication is a person present in your life, but with whom you have no relationship whatsoever. Communication with this person will benefit you even less than weak communication.
Depending on the type of application that you are creating, it may be useful to design it in such a way that people will tend to create weak ties with those people whom they hardly know. It is likely that they will give them preference even more than strong ties. But it is important to distinguish between weak and absent communication. In social networks, such as MySpace and Facebook, where the value of each account actually consists of the value of connections, this difference is quite significant. But the fact that you can establish or search for any type of connection has a significant impact on the growth of such networks.
According to the Granovetter theory, the visual picture of weak links will be significant. In LinkedIn, you can look at all the connections depending on the proximity circle, but there are no indicators of whether these connections are strong, weak, or missing. LinkedIn has another problem: it’s difficult to establish weak links themselves. You often need to ask a mutual friend to make such a connection. Without a doubt, LinkedIn in this case retreats to the background compared to other social services!
For further consideration of the mechanisms of the social network, it is necessary to introduce the concepts of centering and dependence of one node on the rest. In this connection, an article by Linton C. Freeman on centering in social networks will be useful (Freeman, 1979). Freeman investigated how the “centering graph” depends on the differences in the choice of centering points. He also identified three competing principles related to the definition of centering, namely: node rank, control, and independence.
The rank of a node depends on the number of nodes connected to the data. Simply put, it is equal to the number of your friends in the social network. The more friends you have, the more important you are.
Node control implies its effect on communications between other nodes. For example, if hundreds of friends are connected with each other only through you, your centering is very large. In this case, you are the node that controls the flow of information.
Finally, independence means that a node has close ties with all the nodes in question, so it is practically independent of each particular node and is not controlled by anyone. In general, you can reach the maximum number of people using the minimum number of connections, regardless of any particular nodes.
Fig. 1. Group center
Social networks are basic tools with which users constantly monitor and expand their network of contacts. Therefore, most of these sites depict their height, using the rank of nodes as the main criterion. However, control and independence are also important factors. For example, a user who controls the flow of information may be more important than one who has many friends in this network. The centrality also shows which resource participants are most useful or have the most connections, reflecting their greater competence as information sources.
The principles of knot structure, strength of connections and centrality can be applied to existing social networks on the Internet. A good example is the comparative study conducted by Kumar, Novak and Tomkins (2006). They compared two social networks, Flickr and Yahoo 360, which at the time had a total of five million users. The researchers noted that social networks are developing according to the standard scenario, namely: first, rapid growth, then a certain period of decline, after it slow, but stable growth. Kumar, Novak and Tomkins also noted that social activity is of three types:
Fig. 2a Ginger group denotes the central part. The periphery consisting of separate small groups is marked in blue. Lone gray marked
Analysis of nodes for such networks showed that approximately half of the nodes are located outside the large center, in which the main centering is concentrated. For this centering, the definition of “control” was used. The study also noted a large number of "stars" in the periphery, each of which represents a mini social network. Such groups usually have one leader, who can be considered as a centering point, and the rest - as his satellites, connected with the central leader, but not with each other. In the analysis of Kumar, Novak and Tomkins, one third of users on Flickr belong to such groups and about ten percent on Yahoo! 360.
It is also worth noting that the greatest growth occurs precisely on the periphery, where active users are pushing their friends to join their network. This sub-network may eventually join the central part. After such a merger, the importance of such active users is reduced. Even if they decide to leave the community, other users will still remain on the social network.
Fig. 2b. Connections arise between the center and the peripheral group.
Fig. 2c. The peripheral group joins the center.
What conclusions can be drawn from this? When designing your network, you need to consider that most of it will be located outside the central part. Essentially, a social network is represented by thousands of sub-networks. The more opportunities you give these networks for development, the greater will be the overall growth. Social networks, in fact, are virtual ghettos. Those networks that encourage such ghettos (for example, MySpace and Facebook) grow the fastest. In Ning, where you can create your own social network, as well as join others, you are best understood and used this concept.
Most social networks grow on the enthusiasm of the first users who transfer their contacts from the real world to the Internet and act as “stars”. But it is also important to consider the development of social networks based on the internal activity of users. Backstrom, Huttenlocher and Kleinberg (2006) analyzed the creation of groups in large social networks. They used data from LiveJournal about ten million users and DBLP, a database of co-authors in publications from conferences, to study how the growth of communities is related to the social networks underlying them. They showed that the user has more incentives to join a social network if his friends are already connected to each other in it. Several closely related friends in the social network increase the credibility of it. For those of us who are active users of the social network, this seems already obvious.
The article has consistently shown that the main growth is due to the large center, in which the nodes have the greatest centering. Emphasizing the importance of a large center, Backstrom, Huttenlocher and Kleinberg confirmed the theory of Kumar et al. (2006) In their article, a very important question arises: “When a particular node learns about the behavior of its neighbors, what conditions and connections within the network make it accept this behavior? "
Another group of researchers who studied the DBLP database were Cai et al. (2006) They emphasized the fact that each node belongs to several different social networks, each of which affects the general patterns of group formation, development and information exchange in the network. As a result, they concluded that each network cannot be studied independently, but only in the context of other existing networks. After all, it is possible that this user behavior when he leaves one social network due to the activity of other users in other social networks. All this raises a question equally important for developers: “Do you know how much activity in your social network depends on the activity of other social networks?”
Of particular interest is the fact that now Google lab is spending its strength on the development of Social Stream. This project, in theory, should become a meta-social network in which various social networks will be consolidated. Social Stream, being developed in close collaboration with Carnegie Mellon University, is now in a closed beta state. The question that social network architects should think about is the following: “If you get access to managing your activity in different social networks through a single interface, how will this affect your network preferences?”
It is clear that social networks on the Internet will always evolve both due to external influence and because of the activity of users inside them. Butler (2001) emphasized this when he pointed out the following characteristic feature: the size of a social network has a complex impact on the network itself, the more one users succeed, the more other users fail. It was pointed out that it is necessary to achieve a balance between the pros and cons of the size of the network and information activity. And the final question worth considering is: “What type of user activity and in which part of the network (in a large center, on the periphery or among individuals) has the most impact on the network itself?”
In this article, the theoretical principles of the formation of groups and communities, originally proposed by Granovetter and Freeman, are considered in practice. It has been established that socio-technical systems should take into account the actions of users, including their ability to make a unique choice, which will greatly affect the further development of the network. As a result, when using the theoretical foundations of social networks when creating a web product, it is worth remembering that it is very difficult to assess the scale of the potential success of such a network.
In the next articles it is planned to consider patterns of information exchange in social networks. The final part will be devoted to some scenarios of workplace organization.
Author's note: regardless of the complexity of the article, it should be considered only as a starting point when studying the topic touched from an academic point of view.
Thanks to everyone who read the article. I will be glad to your comments. Thanks for attention.
Web Optimizator: checking the speed of loading sites
"The principle of leverage" in action
Users suffer from an excess of information: on any issue, much more information is now available than a person is able to perceive. As a result, people rely on the knowledge of other people. Among the typical questions “how?”, “What?” And “why?”, The key question is “who?” (And, accordingly, the answer to it). This entails the need to meet and maintain contact with people who can help in each case.
')
In the context of the stated problem, it is also important to understand the education, development and use of social networks on the Internet. A social network is a “group of people (or organizations, or other social units) whose connections are social, namely friendship, collaboration, or the exchange of information” (Garton et al., 1997). Although the Internet is the main source of redundant information, it also allows the user to effectively manage their social networks and thus receive the information that they need.
This area of knowledge is located at the intersection of such sciences as computer science, sociology and mathematics. Its researchers are studying the use of social networks for public and business purposes using information and communication technology (ICTs) , including the Internet. This article discusses how social networks, by taking advantage of such technologies — especially Internet technologies — are created, developed, and profitable.
Social networks on the Internet are an extensive field for research, and any scientific review will in one way or another be based on a set of academic publications. This article focuses on the latest research on network center ( centrality ), the strength of connections, individuality, trust, activity and benefits. Undoubtedly, the article is somewhat difficult to understand, but it is intended to convey to developers several useful ideas on the implementation of web applications that actively use social networks.
Strength of weak ties
For the first time social networks were investigated in the late 40s of the last century. With the advent of the Internet, online communities and relevant websites, their importance has only increased. Any self-respecting author who claims to be objective, in his article will have to mention the sociologist Mark Granovetter and the mathematician Linton C. Freeman , who wrote the basic materials even before the widespread use of the Internet.
Granovetter (1973) stated that within social networks, weak links are much more important than strong ones. He explained this by the fact that through weak links information “seeps” faster. This, in his opinion, is due to the fact that weak links are “more important for individual users when they are“ infused ”and interact in the community, whereas as a result of strong links a close local group is formed.
In his doctoral thesis, Granovetter showed that people are more likely to find work because of their weak ties, rather than strong ones. Weak ties are links with people whom we hardly know, with whom we don’t share our feelings and do not maintain relationships; but they turn out to be the most useful. This is due to the fact that people with strong connections share mostly the same data or resources, thus they are less useful to each other.
Along with them, Granovetter identified the missing connections (also called “ nodding ”) - such connections lack the emotional component, time, trust and reciprocity - as weak. Someone living with you on the same street, to whom you nod every day, will be just an example of such a connection. Missing communication is a person present in your life, but with whom you have no relationship whatsoever. Communication with this person will benefit you even less than weak communication.
Depending on the type of application that you are creating, it may be useful to design it in such a way that people will tend to create weak ties with those people whom they hardly know. It is likely that they will give them preference even more than strong ties. But it is important to distinguish between weak and absent communication. In social networks, such as MySpace and Facebook, where the value of each account actually consists of the value of connections, this difference is quite significant. But the fact that you can establish or search for any type of connection has a significant impact on the growth of such networks.
According to the Granovetter theory, the visual picture of weak links will be significant. In LinkedIn, you can look at all the connections depending on the proximity circle, but there are no indicators of whether these connections are strong, weak, or missing. LinkedIn has another problem: it’s difficult to establish weak links themselves. You often need to ask a mutual friend to make such a connection. Without a doubt, LinkedIn in this case retreats to the background compared to other social services!
Network centralization
For further consideration of the mechanisms of the social network, it is necessary to introduce the concepts of centering and dependence of one node on the rest. In this connection, an article by Linton C. Freeman on centering in social networks will be useful (Freeman, 1979). Freeman investigated how the “centering graph” depends on the differences in the choice of centering points. He also identified three competing principles related to the definition of centering, namely: node rank, control, and independence.
The rank of a node depends on the number of nodes connected to the data. Simply put, it is equal to the number of your friends in the social network. The more friends you have, the more important you are.
Node control implies its effect on communications between other nodes. For example, if hundreds of friends are connected with each other only through you, your centering is very large. In this case, you are the node that controls the flow of information.
Finally, independence means that a node has close ties with all the nodes in question, so it is practically independent of each particular node and is not controlled by anyone. In general, you can reach the maximum number of people using the minimum number of connections, regardless of any particular nodes.
Fig. 1. Group center- Node rank: C and K have the largest number of nodes with which they are associated.
- Control: D serves as a bridge between a large number of nodes and controls the flow of information.
- Independence: K has a fairly close relationship with the other nodes through several others (I and Q).
Social networks are basic tools with which users constantly monitor and expand their network of contacts. Therefore, most of these sites depict their height, using the rank of nodes as the main criterion. However, control and independence are also important factors. For example, a user who controls the flow of information may be more important than one who has many friends in this network. The centrality also shows which resource participants are most useful or have the most connections, reflecting their greater competence as information sources.
Learn from Flickr and Yahoo
The principles of knot structure, strength of connections and centrality can be applied to existing social networks on the Internet. A good example is the comparative study conducted by Kumar, Novak and Tomkins (2006). They compared two social networks, Flickr and Yahoo 360, which at the time had a total of five million users. The researchers noted that social networks are developing according to the standard scenario, namely: first, rapid growth, then a certain period of decline, after it slow, but stable growth. Kumar, Novak and Tomkins also noted that social activity is of three types:
- Singles who do not have connections that are central in themselves
- Large center, it has a large number of nodes associated with the center and among themselves
- The periphery, it presents a separate group of nodes with links within themselves, but not having links to the rest of the network. They are a kind of single stars. Such groups grow slowly, one user at a time. Over time, they can join a large center.
Fig. 2a Ginger group denotes the central part. The periphery consisting of separate small groups is marked in blue. Lone gray markedAnalysis of nodes for such networks showed that approximately half of the nodes are located outside the large center, in which the main centering is concentrated. For this centering, the definition of “control” was used. The study also noted a large number of "stars" in the periphery, each of which represents a mini social network. Such groups usually have one leader, who can be considered as a centering point, and the rest - as his satellites, connected with the central leader, but not with each other. In the analysis of Kumar, Novak and Tomkins, one third of users on Flickr belong to such groups and about ten percent on Yahoo! 360.
It is also worth noting that the greatest growth occurs precisely on the periphery, where active users are pushing their friends to join their network. This sub-network may eventually join the central part. After such a merger, the importance of such active users is reduced. Even if they decide to leave the community, other users will still remain on the social network.
Fig. 2b. Connections arise between the center and the peripheral group. Fig. 2c. The peripheral group joins the center.What conclusions can be drawn from this? When designing your network, you need to consider that most of it will be located outside the central part. Essentially, a social network is represented by thousands of sub-networks. The more opportunities you give these networks for development, the greater will be the overall growth. Social networks, in fact, are virtual ghettos. Those networks that encourage such ghettos (for example, MySpace and Facebook) grow the fastest. In Ning, where you can create your own social network, as well as join others, you are best understood and used this concept.
Live Journal, DBLP and Adaptation Behavior
Most social networks grow on the enthusiasm of the first users who transfer their contacts from the real world to the Internet and act as “stars”. But it is also important to consider the development of social networks based on the internal activity of users. Backstrom, Huttenlocher and Kleinberg (2006) analyzed the creation of groups in large social networks. They used data from LiveJournal about ten million users and DBLP, a database of co-authors in publications from conferences, to study how the growth of communities is related to the social networks underlying them. They showed that the user has more incentives to join a social network if his friends are already connected to each other in it. Several closely related friends in the social network increase the credibility of it. For those of us who are active users of the social network, this seems already obvious.
The article has consistently shown that the main growth is due to the large center, in which the nodes have the greatest centering. Emphasizing the importance of a large center, Backstrom, Huttenlocher and Kleinberg confirmed the theory of Kumar et al. (2006) In their article, a very important question arises: “When a particular node learns about the behavior of its neighbors, what conditions and connections within the network make it accept this behavior? "
Another group of researchers who studied the DBLP database were Cai et al. (2006) They emphasized the fact that each node belongs to several different social networks, each of which affects the general patterns of group formation, development and information exchange in the network. As a result, they concluded that each network cannot be studied independently, but only in the context of other existing networks. After all, it is possible that this user behavior when he leaves one social network due to the activity of other users in other social networks. All this raises a question equally important for developers: “Do you know how much activity in your social network depends on the activity of other social networks?”
Of particular interest is the fact that now Google lab is spending its strength on the development of Social Stream. This project, in theory, should become a meta-social network in which various social networks will be consolidated. Social Stream, being developed in close collaboration with Carnegie Mellon University, is now in a closed beta state. The question that social network architects should think about is the following: “If you get access to managing your activity in different social networks through a single interface, how will this affect your network preferences?”
It is clear that social networks on the Internet will always evolve both due to external influence and because of the activity of users inside them. Butler (2001) emphasized this when he pointed out the following characteristic feature: the size of a social network has a complex impact on the network itself, the more one users succeed, the more other users fail. It was pointed out that it is necessary to achieve a balance between the pros and cons of the size of the network and information activity. And the final question worth considering is: “What type of user activity and in which part of the network (in a large center, on the periphery or among individuals) has the most impact on the network itself?”
Conclusion
In this article, the theoretical principles of the formation of groups and communities, originally proposed by Granovetter and Freeman, are considered in practice. It has been established that socio-technical systems should take into account the actions of users, including their ability to make a unique choice, which will greatly affect the further development of the network. As a result, when using the theoretical foundations of social networks when creating a web product, it is worth remembering that it is very difficult to assess the scale of the potential success of such a network.
In the next articles it is planned to consider patterns of information exchange in social networks. The final part will be devoted to some scenarios of workplace organization.
Author's note: regardless of the complexity of the article, it should be considered only as a starting point when studying the topic touched from an academic point of view.
Thanks to everyone who read the article. I will be glad to your comments. Thanks for attention.
Web Optimizator: checking the speed of loading sites
Source: https://habr.com/ru/post/31247/
All Articles