Let Twitter test the authenticity of information

Lead:Dataminr founder and CEO Teide Belly (Ted Bailey) has published an article in the US Technology blog Gigaom, which says that many people feel that Twitter information is low in authenticity and unreliable, but if users try to use a lot of Twitter information is assembled,and Twitter will be a new and important way to confirm the authenticity of information than traditional news sources.

The following is a full translation of the article:

When a new day comes, the first thing people do is to broadcast news on Twitter. In the development of news events, such behavior has spawned a new real-time global Witness, which can be called "Sensor network".

However, in the changing process of news, the verification process of information brings challenges. And these news sources are unique, and each of them uses Twitter to broadcast what they see and hear, at the right time and in the right place. They're just ordinary people with Twitter accounts, not usually known to the public, but now they're on the front line Broadcasting news events.

This feature of Twitter raises questions about the reliability of news production. However, many people have overlooked this point: the early eyewitness release of the Twitter message brought together data patterns that formed the most accurate perspective for verifying the reliability of information.

The reason is simple: Twitter itself is the most effective way to validate Twitter messages.

This week, a devastating explosion in Harlem, New York, killed several people and plunged the whole of New York into deep sorrow. After the incident, viewing the Twitter data enabled us to see information that had been difficult to find. In the Dataminr I founded, we have a huge team of data scientists who have spent years analyzing the variety of data patterns that news events spread on Twitter. We found that Twitter data can be a very interesting model. In the event of an emergency, a small number of Twitter messages can provide enough data for our algorithm to depict the event and determine the correctness, relevance and operability of the relevant information.

A few minutes after the Harlem bombing, aggregated Twitter data showed what was going on. Many people play the role of field agents and sensing networks, and the scene is described in detail when the reporter arrives.

In the Harlem bombings, these Twitter witnesses provided photographs and first-hand information about the puzzle, which came from the same location in a short time. In addition, the location information in the Twitter message, the mode and speed of the message propagation, and the language characteristics of the message itself can quickly reflect the importance and influence of news events. We can know all this before the traditional media rush to the scene.

According to the New York fire station, the explosion took place 9:31 A.M. Eastern time. 12 minutes later, the local media reported it first. 20 minutes later, the mainstream media reported. Within 12 minutes of the explosion, however, there was a lot of detailed information on Twitter, and then more detail.

The following chart shows a Twitter message with position information aggregated by the DATAMINR algorithm within 12 minutes of the explosion. The facts were in the Twitter message before any source confirmed the explosion. The unique pattern of Twitter messages confirms the explosion, and Twitter users see it as "a big deal". The descriptions in these messages provide the raw and important information that the eyewitness experiences in real time. Dataminr observed this pattern and gave a hint within 200 seconds.

Some people say that "the information on Twitter is unreliable" because they ignore the most unique attributes of Twitter's data: the ability to license itself. It is difficult to prove the authenticity of a message when you see a single Twitter message that is unfamiliar to the source. But for breaking news events, the most accurate data description of the event can be obtained by assembling a large number of Twitter.

Can Twitter be a stand-alone way to validate all information? Not all of this is true, but the traditional news system is equally imperfect. All in all, Twitter should be a key and irreplaceable part of the process when it comes to validating breaking news. (Lili)