Published on March 6, 2014
Visualizing Social Media Big Data A big mountain of Social Buzz hiding invaluable insights March 2014 Every global Company now has software to track and analyze the conversations around their brand and competitors. Typical reports vary little and may include sentiment analysis, volume of buzz, top terms & topics, most used social platforms and influentials count. At E.life we have been doing this for the last 10 years and eventually we realized we could do a lot more simply because of one single fact: the available social data has just exploded in the last 36 months. We started discussing alternatives to 1) handle this big data (a whole new challenge for brands since we used to talk about thousands and now we are talking about millions of posts) 2) create effective ways to discover valuable insights from those millions of posts we are gathering every day. To handle the big data explosion we expanded our cloud infrastructure using Amazon services and went from structured databases to text based and unstructured data processing systems. This gave us the possibility of querying and cross referencing millions of social media items in seconds. That was the easy part and involved a few brilliant E.life engineers. One of the first applications of our new infrastructure was done for the biggest Brazilian TV Network, “Rede Globo”. We now process the buzz of all their TV Shows, including the famous soap operas (novelas) and the Brazilian franchise of the music show “The Voice” handling circa 10 million items each month. 01
Managing the process of insights discovery was the next challenge we began tapping into. It soon proved to be an ongoing job, since the range of our clients varies from Oil and Energy to manufactured goods, retail and airlines among others. The most important task we posed ourselves was to ask a kind of meta-question for each client, namely: Which Questions should we ask the Data*? We came up with a simple framework we call the 4 Ps of insights, which guides the search for these questions. We also turned our focus from only performing brand monitoring to including the whole Consumer Universe of a specific brand. We defined the 4 Ps as follows: the general Preferences of a consumer (she likes yoga, watches old American comedy movies and loves Family Guy), the Pricing sensitivity (he likes fast food because it is cheap, he plays golf and does not care if it involves over expensive gear), the Places she goes to (Gym, Starbucks and the local Mall) and what kind of People these consumers are (he is a father, she is a runner, he works at a Bank, she is a Journalist). Amazing as it seems we were able to find a great deal of instances of those Ps by looking at the social big data. We also designed a lego-style dashboard which allows us to quickly visualize the Ps extracted from millions of items on a time interval. Below (figure 1) we show an example of 3 Legopieces (or widgets in the geek jargon) of one such dashboard sitting on top of hundreds of thousands of mentions and digital check-ins at Shopping Malls in Sao Paulo. *An ongoing joke probably from the heroic beginnings of data mining theory in the 70´s was: if you press the data enough it will eventually confess. Figure 1 – three widgets from our Check-in at Malls dashboard 02
The First widget shows the total volume of tweets made in Shopping Malls in Sao Paulo in one day and more importantly how many people had done these check-ins. The second widget shows the ranking of Malls according to volume of check-ins (which we think is a rather useful metric for Mall marketing directors) and finally the last widget extracts from the Bio description how consumers define themselves: Journalists, Ad people and Designers form the top 3. But things get a lot more interesting! When we cross-referenced this data with other categories we were able to figure out even more about the consumers checking in at Malls. Below we show two other widgets we built on the Shopping Malls dashboard. Figure 2 – Related Stores and TV Shows widgets The related stores widget shows the most cited retail chains mentioned by our sample of consumers. We can also filter the data for a specific Mall, giving marketing planners a chance to figure out which chains will appeal more to their customers. Finally the widget “top related shows” tells us a bit about the TV shows consumers most mention on their favorite social platform. Again this information can guide media buyers to optimize their TV, Youtube and Netflix campaigns. gives access to a world of information that traditional market research cannot begin to dream of. A complete dashboard might also include attributes such as gender, musical preferences, fast food or fashion chains, movies, hobbies, sport fans etc. The technology and the data are there, your imagination is the limit. The “Check-ins at Malls” dashboard is just one of infinite possibilities. Social big data 03
Interested in a Social Big Data dashboard for your company? Send us an email at email@example.com and we’ll be very excited to help you to discover incredible insights for your business. Jairson Vitorino, PhD CTO of E.life @jvitorino www.elifemonitor.com/us www.elife.com.br/home_uk
Small deck used during Use Case roundtable at JiveWorld 2014. On each on the 14 ta...
This 30 minute presentation was given at the 2014 Rochester Young Professionals En...
Visualizing social media data to empower individuals. We all know that social media generates huge amounts of data; the explosive growth of social media is ...
Visualizing Big Data: Social Network Analysis ... driven by social media and information visualization. ... Today’s Big Data mining field is prodigious.
Social Media und Big Data? Eine Beziehungsgeschichte. von visualbablog | Aug 15, 2014 | Business Intelligence, Dashboard, Datenvisualisierung | 0 Kommentare
research on mapping and visualizing digital and digitized data ... memory in the Balkans and data ... and Social Sciences at ...
Visualizing Communication on Social Media Making Big Data Accessible Karissa McKelvey, Alex Rudnick, Michael D. Conover, Filippo Menczer Center for Complex ...
Was diese interaktive Visualisierung ermöglicht, ist eine Auswertung von „Big Data“. Visualisierungen helfen uns dabei, diese Menge an Daten besser zu ...
Using 3D Printing to Visualize Social Media Big Data ... have difficulty in visualizing the often ... Working with big data usually implies working ...
Visualizing Social Media Analysis. ... There’s a lot of hope attached to idea that the use of big data-drivers like social media can create important ...
Big Data Visualization: ... From visualizing national marketing campaigns to mining and ... and unstructured data sources such as social media and sensor data.