This blog post will introduce you to the field of network theory and the study of social networks. It will explore some types of network characteristics and network theory techniques and discuss them in relation to social networks on EventHive. Because no network analysis has been performed, this blog post remains both speculative and descriptive. Instead, this post suggests ways in which network theory and network characteristics might apply to the EventHive platform.
A distinctive feature of EventHive is that offline and online networks collide and networks form both in the offline (the actual event itself) and online (EventHive) space of the event. In the final section of this post, this problematic will be discussed further and will point to work that attempts to compute this phenomenon.
Who and What Are the Nodes and Links?
Networks are made up of nodes and links and networks represent relationships amongst objects. A node is the object, and a link is the relationship. When looking at EventHive, nodes might represent people and event organisers.
Strong and Weak Ties
The relationships between objects, or the links between nodes, can vary in strength. The strength of links is often described in terms of being a strong or weak tie. A stronger tie is likely to reflect a higher frequency of interaction or interaction type. As Easley and Klienberg advise, what constitutes ‘strong’ and ‘weak’ in a network-theory analysis needs defining from the start [1].
On EventHive, a strong tie would emerge between people if they decided to chat online or form a group conversation, rather than implicitly – or weakly – being connected because they have registered to attend the same event.
Degree Distribution
Degree distribution refers to the total number of links a node has; the higher the number of links, the higher the ‘degree distribution.’ From this, Wang and Chen suggest one can infer that the higher the number of links the more ‘important’ the node is [2]. Statistical operations can be performed using the degree distribution, such as calculating the correlation between some explanatory factor and the value of the degree distribution.
When analysing EventHive, the degree distribution could be used to assess an event’s popularity. Let’s assume that events and people are both nodes in a network, and registering to attend an event forms a link between a person node and an event node. Therefore high degree distributions of an event node shows that more people that are ‘interested’ in it and likely to attend that event. In turn, we could infer that this event is more ‘popular.’
Clustering Co-Efficient
The clustering co-efficient is a statistic used to assess the probability that two randomly selected nodes (X,Y) and linked with node (Z). It can be used to observe clustering, or grouping, of nodes. Large-scale networks have a tendency towards clustering and there is often a ‘real-world’ explanation behind its occurrence [3].
In an analysis of EventHive, it would be interesting to analyse network clustering amongst attendees of an event. It would be interesting to explore whether there are any relationships amongst event attendees that extend beyond the fact that all of these people are attending the same event. This analysis would involve isolating clusters in event attendees networks and investigating what factor, such as common interest, friendship etc., links these groups of nodes.
Event-Based Social Networks: Combining Offline and Online Networks
In this blog post so far, I have tried to focus the discussion of network theory and EventHive to the online user experience i.e. how people and event organisers connect on our event platform. However, events are not online social activities and they are very much grounded in the offline; an online application is only an extension of the ‘real’ experience of going to an event. Consequently, we have to find some way of grasping how online and offline networks collide and come together when some uses an online social network at a live event.
The Application of Network Theory to “EBSNs”
Liu et al describe these types of online/offline social networks as Event-Based Social Networks (EBSN), and highlight how event-based online social networking services contain both online/offline social networks and online/offline social interactions [4]. In this research, EBSNs (MeetUp.com and Gowalla) have been analysed using network-theory techniques, including degree distributions and clustering coefficients.
Degree Distribution
It Liu et al’s paper, the degree distribution is used to independently compare the network properties of offline and online event-based social network, as well as consider how these two networks interact with each other. Lui et al found that that the online social network was denser than the offline social network, meaning more people connect online than at social events. In addition, the analysis found a correlation between the degree distribution of these two networks, meaning each network has a positive effect on each other – more online connections equals more offline connections and vice-versa.
Clustering
Liu et al define distinct online or offline social networks as ‘homogenous’ networks, and combined offline/online event-based social networks as ‘heterogeneous’ networks. In order to analyse clustering, the researchers partitioned the social networks, meaning they isolated node clusters. In their subsequent analysis of these clusters, Liu et al found that in the MeetUp.com online social networks there was more clustering around similar interests, rather than clustering based on the participation in the same offline event.
Summary
This blog post has attempted to introduce you to the field of network theory. Jack has introduced some basic concepts and techniques used in network theory to analyse social networks. Throughout, this post has attempted to relate these concepts and techniques to EventHive and the various social networks that operate on this platform. The final section of this post discussed attempts that have been made to analyse heterogeneous offline/online social networks, which are evident on services such as MeetUp.com as well as EventHive.
References
[1] Easley, D. and Kleinberg, J. Networks Crowds and Markets. Cambridge University Press, 2010.
[2] Wang, X. F., & Guanrong Chen Circuits and Systems Magazine, I. (n.d.). Complex networks: small-world, scale-free and beyond. Circuits and Systems Magazine, IEEE, 3(1).
[3] Ibid.
[4] Xingjie Liu, Qi He, Yuanyuan Tian, Wang-Chien Lee, John McPherson, and Jiawei Han. 2012. Event-based social networks: linking the online and offline social worlds. In Proceedings of the 18th ACM SIGKDD international conference on Knowledge discovery and data mining (KDD ’12). ACM, New York, NY, USA, 1032-1040.
No comments yet.