The human face of large data

The so-called "large data human faces" (the Human face of the great data,hfobd), please do not literally, misinterpreted as having a close connection with the face recognition in video or image-although, the initiator of the "Big Data Human Face" project Rick Smolland (Rick Smolan) is a famous photographer.

The project is a global endeavor aimed at introducing humanity to the revolutionary role of our people in life, learning, governance, work, and play. The project will showcase the changes that big data has brought to our world through simple, humane stories and images, and indicate some of the ways in which it will affect us in the future. The project also provides large data as a cornerstone of the activity itself: it brings together millions of people around the world to serve as a "human sensor" for the day, providing information about their thoughts, actions, opinions and experiences within a 24-hour period in 2012. The project has six main components: human sensor components (smartphone applications), large data visual records in action (printed hardcover and ebook), "Command and Control Center" (burritos controls) experience, large data tracker, data visualization toolkit, media and social media promotion ...

"Big data human face" Smartphone application, initially only Andorid English version, but has been seen strong social attributes

Taking "human sensor components" for example, Rick Smolland and his team developed a "Big data human Face" Smartphone application (free downloads of iOS and Android versions in five languages) to "Measure our world." I installed the app in an English version of Android, and answered N more questions such as "What do you think happens after death", "raised/Never had a pet"? In Singapore's "Command and Control Center" experience, the results of analysis from millions of participants were presented, such as:

In the "Family" section, there are "parents from childhood to me (severe/tolerant)" and "I compare (pessimistic/optimistic)" and so on. Analysis found that from a strict parents, the personality will be more pessimistic when they grow up. Uh ... This, how to say? The exploration of this project is worthy of encouragement, but the analysis of the results, it seems that only confirms our common sense, and does not reflect the most important feature of the--value data.

"Big Data Human Face" Android platform application, shows that at the end of September more than 100,000 people participated in the survey, a week later with the promotion of the activities of nearly 2 million, now more than 3 million

As the main sponsor of the "Big Data human Face," Steve Leonard, senior vice president of EMC, said when he mentioned the project, "How much do you think the YouTube 24-hour video, every 60 seconds, is being uploaded as part of human activity?" But the Greenplum division, which is the flagship of EMC's Big data analysis, shows up in Singapore's command and control center, mainly on data visualization that works with Twitter. To achieve this goal, EMC has set up 1000 nodes of the Greenplum cluster in Las Vegas, gathering Twitter information and analyzing it. In EMC's words, the amount of information is much like a fire hose (firehose).

At first glance, it's much easier to analyze the text content of Twitter than the "face recognition" in a picture or video. However, it is too difficult for software to identify words such as "Romney" or "Obama" easily, and to judge emotions and attitudes from the context of human language.

As a friend of mine said, "I love Obama" and "I love Obama", and how different the attitude, in software analysis, is definitely a problem. English is not good to go there, so in the Twitter example shown in Greenplum, it is also focused on digging voter tweets with the Obama or Romney relationship, who is talking more, but not as a basis for support ratings.

After the election, an article in Time magazine analysing the new data-analysis strategy used by the Obama campaign in this election has warmed the tide of big data. 08 social Networking, the 12 use of large data, Obama's two campaign to perfect the interpretation of the "Times", it is a tidal burst. This time, some say, socializing is the front desk for Obama to get the public opinion, and in the background, the big numbers support Obama's various campaign strategies and decide what social platforms he should go to. However, from this article, it is difficult to see the large data methods and social networks, the depth of the content of the excavation, the main play seems to still belong to the telephone, e-mail and other relative "traditional" means.

The statistics on Twitter are consistent with the information released after the election.

Some people may ask, "Big Data" topic, why catch the social network do not put? This is because social networks generate a stream of text, pictures, video messages, volume and produced (diversity) definitions that fit large data, enough complexity (complex), It also requires velocity (fast) processing, but as the previous analysis, people directly generated (such as Twitter and Facebook text) or contain complex human activities (such as photos, video) data, the machine is difficult to judge, by the existing technology constraints, the current can produce value (value) and relatively limited.

By contrast, various sensors collect, the record of simple information (such as location information, not video, image), which conforms to the large data recognized three V-a-c definition, but also relatively easy to deal with the analysis, has shown great value, which has splunk success and a variety of user behavior analysis of the attempt to prove.

These data are more widely sourced (Volume) and more complex (complexity) than the business-critical transaction data of the past, but can still be placed in a variety of databases or data warehouses, with more rapid (Velocity) processing of emerging technologies (produced) , it is difficult to cover the traditional trading system.

In other words, big is not secondary, "Big data" brings us the inspiration is to pay attention to the traditional enterprise transaction database, all other data value-especially the so-called "passive data" that many sensors automatically collect. There may be a lot of meaningless spam in some types of data, but the point is that each type of data has to consider how to organize it effectively.

These data, of course, include more complex social data. Perhaps there is not enough value at the moment, but in the big data rise, multifaceted, the manufacturers are busy to seize the site, in the guarantee of their own interests at the same time, for future growth in advance layout. This year's Oracle Open World has a page of presentations that almost equate large data with social data, at the bottom of the data pyramid relative to Oracle's core databases and data warehouses. And Steve Leonard's passage was intriguing:

"Think of all the information that is generated." Think again. Usually 100 times times the amount of information is only in the transmission, not save and protect, just flow through the system. Each of the information that is actually saved is hundreds of times times as much as the amount of content. So all of us, every day, are producing huge digital footprints, or digital shadows. This is what people do every day as individuals. ”

What do you think? Anyway, I feel like, what EMC is saying is that social data is not fully utilized right now, but it must be preserved first and will be available in the future ... Well, that's true, and it's good for you and for EMC. From this perspective, it is also possible to explain partly why the value of large data is recognized as an analysis, but the store is one of the most enthusiastic.

Whether you accept the notion of big data, or how long this boom lasts, the "alternative" battlefield surrounding the data story is open. Manufacturers from their own starting point to tell the story, competition is who can impress the customer's heart, the story is to say success. And then, you know ...