The Under-Represented Perspective of Our Data Trails
The Internet has forever changed our lives by the way we interact, socialize, gather data, read the news, and do business. The basis for interaction has morphed drastically through the development of social media platform such as MySpace, Facebook, Twitter, LinkedIn, and the list keeps growing from there. The mentioned platforms gather more than 1.5 billion users, and with such an information gold mine would develop interest across all the job spectrum, from sociologists to CIA agents, and even terrorists.
The study takes course on Twitter, the 140-character platform that has managed to topple tyrant regimes and inform the world within seconds on crisis such as the Haiti earthquake, Bin Laden’s death, and the list goes on. Twitter, since 2011, has become one of most used ubiquitous news-funneling source in the history of mankind, and there’s enough room to grow beyond.
The Twitter platform engages its users by an engulfing desire to influence and the desire to be popular. A tweet goes into a void space, manically wishing to be retweeted, to be replied to, and to be followed. The numbers of followers and tweets sent are an addiction indicator. The more influence a user craves, the more effort he invests into increasing this number. In other words, Twitter is a total chaos of 140-character chunks of noise with a couple of intellectual notions here and there.
My project targets this phenomenon. By human nature, anything un-interesting or out of context would be ignored. That same principle works otherwise when it comes to popular posts, news that go ‘viral’, which is a term often used for an online event that spreads rapidly, just like a virus, people get to retweet it, talk more about it, and interact with the one who posted.
Social Collider is a tool that measures just that. Over a defined time frame, user tweets are gathered, analyzed, and graphed. Tweets that are not shared by anyone would just connect to the next item in the stream. The tweets that did resonate, however, will spin and connect horizontally to the users or topics that correlates with them directly or in terms of related content.
What you’ll see:
The attached video will start by displaying all tweets in a two-dimensional table showing the date and corresponding user or hashtag relating to my twitter account, @nicocohayek. My profile comprises of the following inputs:
- Following : 1,037 Twitter Accounts
- Followers : 635 Twitter Accounts
- Number of Tweets : 6,118
- Joined : October 20, 2010
Upon input, the tool gathers the following:
- Tweets during a time-frame (for simplicity, we’ll set it to 1 week)
- Hashtags used in a tweet, commonly used to categorize tweets.
(Eg. Today’s #News covers the launch of India’s long-range missile)
- Twitter users who interacted with the tweet
- Twitter users with relevant content based on the common hashtags used.
Once the tweets are plotted on the graph, the solver starts correlating the tweets with its users, showing with a color-curved line the conversation path that occurred from the tweet. If the post doesn’t resonate, it connects to the next tweet in the stream. If it does interact with another user, it will create a spin and horizontally link itself with users who related to it, directly or through the common content.
The render process took a total of 41 minutes. The longer the time frame and wider the influence network, the greater time it takes to analyze all the data gathered. Initial samples from different accounts took more than 16 hours to complete the process, while low-profile accounts took a matter of minutes to finish.
The result obtained was a combination of clusters circulating across the center of the graph. The reason is due to the concentration of tweets pertaining to the following queries.
The reason stems from the experience gained through time that led to more content-oriented approach focused on user behavior and an increased network influence. The term #Infographic has a cluster of its own due to the daily posting of this hashtag in a specified time, which created a trendy topic and an upcoming event among Twitter followers/
It is easy to spot the disregarded tweets and the clusters formed around a user or a hashtag. White clusters are generally online trends that can be monitored. The tweets with the most ‘colored’ are the ones that show the degree of influence of a certain account, and reveals the degree of influence and traces the conversation progression across dozens of other users, all in a limited timeframe.
In a nutshell, the users interacting with my tweets created a network with shared content. In a world full of noise, this network has developed trust and reliability on certain users that render instant feedback. This phenomenon applies to the little people as it applies to celebrities. The data in the graph are no longer under-represented, but traced through their journey from node to node in a network of influenced and influencers alike.
The recording of the whole process is visualized in this YouTube video. For best quality, please set to 720p.