web analytics

Bill Bennett


When news goes molecular

I’m excited by South China Morning Post editor-in-chief Reg Chua’s effort to find the new molecular building blocks of journalism.

In the print era, the news story was the basic block. Chua says stories are less valuable in a digital age and daily news stories have even less worth when readers come back to them at a later date. Returning to old stories is now easier thanks to online newspaper archives and search engines.

Chua describes how software tools cleverly pull atoms of news (facts) from sources then knits them back together to form Molecules of News. In effect this means mining raw data for useful information.

In some ways this isn’t too different from the way journalists research sources when writing news.

As every journalist knows, much of the raw data collected in daily news gathering never makes it into news stories.

Chua says the news industry misses the value locked in that data. He thinks the challenge is for news organisations to find ways to turn this into money.

Clearly one approach is to chain atoms and molecules of news together in ways that make it easier to extract information. This means thinking about data structures, not news stories. This could involve writing reports (or atoms, or molecules) directly in to a pre-built data structure.

Chua’s last idea – suggested in a reply to my comment on his post – is the part I find exciting. I’m going to make developing a working news data structure my background project for the year.

I’d be interested to hear any thoughts on how this can work for a freelance journalist.



4 thoughts on “When news goes molecular

  1. At first glance this is about supplying meta data about news rather than just the story, meta data supplies information that gives context and verification to the news. Meta data could also include information about how that news has been received. This is, as has been alluded to in Chua’s article, more valuable than the story itself.
    However the problem from the New Zealand prospective as I see it is that there is a very small periodic table to work with.

  2. Bill, thanks for the shout-out; I’d definitely be interested to see what you come up with for a news data structure. I recall my days covering foreign exchange trading for Reuters, and I’d say 95% of those stories could (and generally did) fall into a simple template.

    Phillip, just wanted to make a quick comment on your comment: It’s not so much about supplying metadata, I think. The weakness (and strength) of metadata is that it often applies to a story as a whole, rather than to a specific element of it, making it hard to strip out the part you really want. You can, it’s true, micro-tag parts of stories, but I fear that without a more formal structure, it’s hard to get past some of the semantic issues in truly building a new product from those atoms.

    And it may well be that the periodic table in NZ is small – although the land does make up for it in natural beauty. But I’d suggest that that’s why targeted projects – a bit like our WhoRunsHK site (https://structureofnews.wordpress.com/2010/12/16/who-runs-hong-kong/) – can be valuable, by making a virtue out of a limited universe.

    Best, Reg

  3. Love it! Very excited by the opportunities this kind of thinking opens up and the potential for new journalism products.

    There are too many mainstream media outlets in this country rewriting press releases, rejigging the same soundbites, at best adding little value to current news and at worst creating misinformation and causing confusion.

    Politifact (cited by Reg) is an excellent example of being able to break things down, providing solid info without the spin. I think there are excellent opportunities for freelancers to develop similar products for targeted, niche information in NZ. For a success story in NZ, I’d look at Consumer.org.nz

  4. are you guys familiar with XML database? How data is laid out in XML

    Keyword based structure, for example.

    australia storm
    no one's killed, blah blah blah blah
    no one's killed
    there has been  2 million loss .... (you can use programmes to search & grab the sentence containing the numbers)

    I’m a bit off topic here may be, if you don’t understand, let me know..

    It’s called a data miner..strip news or complicated piece of information into a few sentences..

    Not sure if it’s done before, but I’m sure someone had that idea.

Comments are closed.

%d bloggers like this: