Fri - 24.11.2017

Politico and Facebook team up to use data for political journalism

Politico and Facebook team up to use data for political journalism

As the Republican primaries heat up in the US, news organizations everywhere are trying to predict the winner. But while none of them have a crystal ball, Politico does have a new source of knowledge at its disposal: Facebook.

Last Thursday Politico and Facebook announced they were partnering up to measure user opinion of Republican candidates. A Facebook data team is using an automated process to measure how often users are mentioning, sharing and linking to GOP candidates and whether these mentions are positive or negative. The results are handed onto Politico, whose journalists analyse and publish them. The first set of data is already out.

The project has been greeted by many as great opportunity to paint a detailed picture of voter opinion. Mashable writes that the vast number of Facebook users and the huge amount of information that Facebook holds on them means that the social network "can become a unique space to conduct survey-based research". Read Write Web writes that "Facebook could be the biggest, most dynamic census of human opinion and interaction in history".

More data, more political journalism, more results to be discussed on Facebook... Everybody wins, right?

Not necessarily. The Politico-Facebook partnership has sparked controversy because it will include an analysis of users' private status updates and messages and, although a machine rather than a human will do the job, the process is still seen by many as a breech of privacy. Read Write Web, for example, which has many positive things to say about the potential for the project, nevertheless laments Facebook's "failure to talk prominently about privacy protections, failure to make this opt-in (or even opt out!) and the inclusion of private messages".

On the other hand, since the analysis looks at the mass data rather than the opinions of individuals and since no humans are reading private messages, some writers dismiss the objections. Blogger Lauren Weinstein, for example, compares the analysis to using looking at healthcare data; it would obviously be a breach of privacy to publish an individual's health records, but so long as anonymity is preserved, analyzing health data on a large scale can be used to spot trends in diseases and save lives.

There is no doubt that big data can be valuable for news organisations and for society - Marshall Kirkpatrick of Read Write Web also points out how it has been used in the US to prove and legislate against racial discrimination in the housing market.

Still, the value of the Facebook data has been called into question. Both Vator News and All Things D have pointed out that, according to another analysis, Ron Paul received more positive mentions on Twitter than any other GOP hopeful but is by no means the most popular candidate; he trailed far behind Mitt Romney in last week's New Hampshire primary. Politico underplays the amount of conversation about Ron Paul in its own data, and one commenter on the Facebook article remarked: "I like how in the Politico link they explain away Paul's continual lead in mentions... what's the point of a graph if you don't take the results to be indicative of popular sentiment?"

There are a other objections to the process: sentiment analysis doesn't necessarily provide the cleanest of data (search for the term "Ron Paul" could throw up random mentions of "Paul" for example) and a machine's judgement of whether a statement is positive or negative doesn't take into account the nuances of users opinions.

Yet while readers may need to take results with a pinch of salt, there's no doubt that the potential for Facebook data mining is huge. Editor-in-chief of Politico, John F. Harris, remarks "Facebook has been instrumental in expanding the political dialogue among voters and we couldn't be more excited about the opportunity to offer our readers a look inside this very telling conversation."

The result could be fascinating not just for people following the election, but for news organisations looking to use big data as well.

Sources: Facebook, Politico, Mashable, Read Write Web, Lauren Weinstein, Vator News, All Things D, The Hill, Presidential Election News


Hannah Vinter


2012-01-16 18:01

Shaping the Future of the News Publishing

© 2015 WAN-IFRA - World Association of Newspapers and News Publishers

Footer Navigation