Twitter has a credibility problem. Fake information and photos proliferate the platform, especially after a natural disaster. For example, Hurricane Sandy photos depicted flooding of the whole East Coast and the Statue of Liberty being felled by a tsunami.
For every problem, there’s a startup. Tweetcred appends a ‘credibility ranking’, rating each tweet from one to seven. The algorithm looks at 45 inputs, including tweet length, if a URL was included, and the number of followers of the tweet source. Tweetcred also learns over time, and users can tweet their own ratings to improve its accuracy.
The Tweetcred score can have a positive impact in for those who Twitter in their professional work. Journalists would find it helpful to concentrate on the most credible sources in breaking news events. Traders might use it as another data point in assessing, say, the impact of Hurricane Sandy on insurance claims.
There remains the underlying issue of Twitter use as a social outlet. Fake pictures went viral because they were the most incredible, not the most credible. Most people didn’t post these fake photos, they simply retweeted them.
It remains to be seen how Twitter addresses the problem of the platform being used for both professional and social use, and how to ensure a positive trait for one use case doesn’t translate into a negative impact for the other.
Read the full article here.
In conjunction with Big Data Week, we compiled a list of the best Big Data articles here.
Leave a Reply