Haga clic en el documento para leerlo a pantalla completa
sábado, 25 de febrero de 2012
martes, 14 de febrero de 2012
The Links of Love
de Facebook Data Team, el Martes, 14 de febrero de 2012
Some people marry their high school sweethearts and live together ‘til death do them part’. Other couples go their separate ways and find love anew. All these collective romantic movements can be used to construct a relationship graph. Relationship graphs (Susan dated Joe who previously dated Jane...) are difficult to maintain on a napkin or a whiteboard because who’s dating whom can quickly become yesterday’s news. But on Facebook, updating a relationship status is as easy as a couple of clicks.
As Valentine’s Day approached, we wondered just how large the Facebook relationship graph might be. So we looked at the romances and splits among people on the site over the past three years.
Since 2009, more than 260 million Facebook users worldwide listed themselves as being “in a relationship” with another person on Facebook. Of these, 65% noted just one person they were in a relationship with. But it is those who dated more than one person who help connect the relationship graph.
When we connect people through their relationship statuses, we find 85 million individuals (32%) are in one giant component, connecting people across the globe. This means that by following the relationship ties from any one person within the component you can reach any other person through relationship ties. For example, if Amy dated Bob who dated Cathy who dated David, then Amy and David are connected.
But before we can conclude that a large part of world is extensively romantically linked, we should note that not everyone uses Facebook's relationship feature to indicate a romantic relationship. Often young women specify that they are “in a relationship” with their BFFs (“best friend forever”). Roughly 20% of all relationships for the 15-and-under crowd are between girls. This number dips to 15% for 18-year-olds and is just 7% for 25-year-olds. With this in mind, we temporarily omitted those ties from our analysis to recalculate the giant component and found the component to still to be 45.4 million individuals strong.
So what exactly does this network look like? It’s difficult to visualize such a large network in its entirety, but we can explore a sliver of it as an example, focusing on US users who were over 18 at the start of the relationship.
We picked a random person, and then picked one person they were in a relationship with. Next we picked someone that the second person had been in a relationship with and so on. Stepping through the network in this manner, we saw long chains of affection.
Visually, we can see that the network is not tightly clustered. In fact, the average of the shortest number of steps to get from any one U.S. user to any other individual is 16.7. This is much higher than the 4.74 steps you’d need to go from any Facebook user to another through friendship, as opposed to romantic, ties.
But remember – we’re only looking at the past three years of data, making our relationship graph quite young. As more relationships blossom, the relationship graph is bound to fill in with new romantic ties. Good luck creating (or reinforcing) one of your own this Valentine’s Day.
martes, 7 de febrero de 2012
Applying Social Network Analysis to Online Communications Networks
Submitted by Natalia Castaneda on Tue, 01/24/2012 - 16:46
By Claire Reinelt, Natalia Castaneda
Looking to increase your reach and influence in the social media space? Social Network Analysis (SNA), a research methodology that focuses on “mapping and measuring relationships and flows between people, groups, organizations, computers, URLs, and other connected information/knowledge entities,” (Orgnet.com ) may be the answer. We recently partnered with the Robert Wood Johnson Foundation to explore how to effectively apply social network analysis to public health online communications strategies, how communications networks operate in Twitter and the blogosphere, and how to identify strategic and influential connections that can be nurtured over time to extend the reach of public health messaging. This was an innovative project that produced detailed and insightful information about how to use SNA to strategize communications campaigns, and we wanted to share some of these insights with the community – including specific recommendations for identifying key messages, influencers, and engagement strategies.
We were especially interested in finding conversations and influencers around the social determinants of health, which according to the World Health Organization, are “the circumstances in which people are born, grow up, live, work, and age, as well as the systems put in place to deal with illness. These circumstances are in turn shaped by a wider set of forces: economics, social policies, and politics.” We were interested in finding conversations that were linking the topic of health with other related topics, such as business, city planning, food and education.
The main questions guiding the exploration included:
• Who are the social media influencers for the relevant topics?
• How can the individuals and groups in public health expand their reach and influence among the influencers?
• How are different public health players positioned in the social media space?
• What are the conversation patterns of the top influencers and hubs?
How We Got Started
Towards the beginning of the project, we spent a lot of time developing a keyword strategy that would allow us to find conversations in Twitter and blogs. Here are some of the strategies we used to develop the keyword strategy:
- Interviewed public health stakeholders and influencers
- Monitored conversation streams to identify hashtags
- Used tools such as Radian6 and SocialMention to test the relevancy and activity of each keyword
- Developed a list of over 40 keywords and hashtags and used it for both the blog and Twitter channels
- Encompass a wide range of topics in our keywords, from mobile health (#mhealth), to issues related to food and health (‘food desert’) and corporate wellness (#co_health). The variety of themes allowed us to find influencers that were leading multiple related conversations, and also to see the connections between people and topics in both channels.
- Leveraged cutting-edge software and technological approaches such as NodeXL and Nexalogy to analyze social media conversations and connections
We developed a series of maps to see where there were relevant conversations happening in Twitter and how those conversations were connected. The following example is a map of the clusters generated by keyword searches relevant to urban health.
In analyzing the map we see that different conversations have different shapes. ‘Complete streets’, ‘walkability’ and ‘smart growth’ are distinct conversations, but they are highly interconnected in this cluster. There are a significant number of yellow squares, which represent Twitter users that are using two or more keywords in their tweets. They are actually weaving together these three keywords or hashtags. Some examples include: @Transportdata, which is an organization that is focused on transportation policy and is bridging between ‘complete streets’ and ‘smart growth’; and another example is @jeromeoppenheim, who is a free agent covering urban policy issues and bridging between ‘complete streets’, ‘smart growth’ and ‘RWJF’.
By contrast, ‘saveplay’ is a highly interconnected conversation with people retweeting and responding to each other, but with many fewer bridgers to other clusters (only two). And ‘healthy food access’ has a very different shape—many more dispersed nodes. This is less a conversation, and more a group broadcasting messages. RWJF seems to be driving this group, along with @naomistarkman (from civileats.com). Also, @Healthyamerica1 has a unique position in this map– the account is bridging between the ‘healthy food access’ and ‘complete streets’ conversations.
Blogs Influencer Network
The following map represents the connections between the top 112 publishers (which sites are linking to which other sites in the dataset):
Three main topical clusters emerged from the blog research: Urban Issues (city planning, urban development, etc.) in blue, the Food and Health (including the food desert discussion) in purple, and the General Health in red. The blogs discussing General Health (red) are highly connected to the blogs discussing the Urban Issues (blue). The General Health cluster has the highest number of influencers, including HealthPopuli.com, HealthAffairs.org, TheHealthcareBlog.com, and the Urban Issues cluster has the second highest, including Good.is, TreeHugger.com and WorldChanging.com. The blogs discussing the Food and Health issues (purple) are not highly connected to the other two main clusters (General Health and Urban). Some sites, such as Civileats.com, are trying to bring the food issues to the forefront of the urban and general health conversations. Also, if you look closely you can see there is an opportunity to strengthen blogging in this space, since there isn't a leader driving the discussions.
Some of the blogs we studied have remarkable positions in the system of blogs that we looked at during the study, such as Wonkroom.thinkprogress.org, which is bridging between the health and urban clusters. However, not a lot of sites are linking to this blog, which represents an opportunity to leverage this site as a connector.
Based on the Twitter and blog maps, we developed some recommendations for those of you who are interested in partnering with social media influencers to increase your reach and the impact of your messages:
The first step is to identify relevant keywords, then identify users. Once you follow them, they are likely to follow back. After you monitor their messages and retweet them, they are more likely to follow you back and even retweet some of your content. So in the meantime, the focus should be on continuing to generate quality content.
Connect with influencers
We identified a number of important "hubs" and "bridgers." We encourage organizations to connect with hubs -- people that are highly connected in the network such as @good and @urbandata; and with bridgers -- people that are active in multiple conversations, such as @Ithealthnonprof, who is discussing issues around ‘rural health’, ‘urban health’ and ‘mobile health’.
Some of the influencers in the blogosphere include: hubs that are actively discussing multiple issues (such as Good.is) and bridgers (such as Wonkroom.thinkprogress.org) that are connecting clusters. Other examples include TreeHugger.com, Futurity.org, Colorlines.com, Civileats.com and Switchboard.nrdc.org.
In Twitter, these are some of the conversations we recommend actively participating in since they seem to be the most active and relevant ones: ‘complete streets’, ‘smart growth’, ‘save play’, ‘food desert’, ‘healthy schools’, ‘co_health’. Co_health is an example of community that is at the intersection of health and business conversation, focused on corporate wellness.
In the blogosphere, there is an opportunity to help strengthen the Food category (since the blogs covering this are not connected to other clusters, so they have low visibility and influence). This would mean discussing this topic actively on your organization’s channels, as well as reaching out to main blogs in that category (asking to contribute guest blogs on relevant topics, etc.), and participating in conversations that are happening in other blogs/communities that we identified. Also, the urban issue area was quite active, particularly conversations around walkability and using bikes.
Participate in relevant events
Participating in relevant events, in this case TED and SXSW (South by Southwest), is a good strategy. By participating in the discussion and using the conference hashtag, you can increase your visibility in that community and attract more followers.
Increase community engagement
In general (both in Twitter and blogs), we recommend increasing the community or audience’s engagement by closing triangles regularly, which means introducing people that are connected to them but not necessarily connected to each other. This will help strengthen the core of the network. According to Nurture website, the definition of network weaving is to "make introductions by describing strengths of each, point to alignment and mutual benefit, name small first step.”
Also, continuing to host webinars and meetings to share ideas and content (even partner with influencers to host some of those meetings), leveraging your organization’s blog and Twitter accounts to ask questions and request feedback on ideas and projects, and to share learnings with your community.
We believe SNA is a highly promising assessment approach for finding influencers in communications networks, but there are many questions that remain, such as
- What is the context for the influence of key players?
- In the process of identifying the key influencers in the health-related blog space we noticed that the blogs and Twitter users vary significantly from niche sites to more established organizational players. It would be interesting to further explore the connection between the types of users, the types of messages, and the types of activities they are engaged in. How do influencers differ in the way they use this media? What are the resources and information key influencers value?
- What is the content of the conversations and interactions?
- Another layer of analysis would focus on understanding the nature of the conversations about the social and economic factors that affect health outcomes –What are the key messages that are being communicated? Are these messages helping to catalyze conversations and interactions across sectors?
What are some other interesting applications of SNA? What other tools and methods have you used to analyze connections in social media channels? We look forward to hearing your questions and ideas!
- Jon R. Katzenbach and Douglas K. Smith, The Wisdom of Teams: Creating the High-Performance Organization (Harvard Business School Press, 1993): More detail on the three types of teams, including “real teams.”
- Jon Katzenbach and Ashley Harshak, “Stop Blaming Your Culture,” s+b, Spring 2011: The next step for an effective top team is reinforcing and building new behaviors that make the most of the existing corporate culture.
- Tim Laseter and Rob Cross, “The Craft of Connection,” s+b, Autumn 2006: How organizational network analysis can improve performance throughout a company.