After looking last night at the way in which algorithms are increasingly influencing our individual writing styles, I came across another article that keeps the language theme going.
With each passing day, we’re generating greater amounts of data collectively on a scale that’s never been seen before. Of course, this leads to all sorts of questions regarding the security and ownership of such data. But it also raises one particular question – who actually has the time to analyse this mountain of data as it grows?
A company called Narrative Science has developed software called Quill that turns data into written text to help to deal with this issue. The company’s only four years old but it’s already being used to report on sports games and company financial results for a range of clients, including Forbes.
As you would expect, those who believe that robots are soon going to be displacing human labour are not the biggest fans of the technology. But whilst the end product is supposedly still recognisable as the output of software as opposed to human creativity, it is sufficiently advanced to do things like vary its writing according to the target audience – so the report of a sporting loss can be written in a less painful way for the beaten team’s supporters, for example.
If the value here still isn’t clear perhaps an example will help. I first came across the company a couple of weeks ago and signed up to use Quill on the analytics that lie behind this website. Given the nature of this site, I rarely spend much time looking at those figures (there’s no real goals for the blog other than having a place to write and to share interesting information). After a few days I received the first automatically-generated report. Written in plain English, it clearly points out the posts that were most/least popular, where people were spending their time and other such useful and, yes, ‘readable’ facts. Whilst I wouldn’t spend much time in Google Analytics, it was no effort at all to flick through this single-page email. There’s no doubt having seen it in action that, in this case, it was definitely more engaging.
There’s another really interesting angle here to me. The company’s had investment from In-Q-Tel, the CIA’s investment division. It’s not hard to imagine that the strongest use case here is actually to shape vast quantities of surveillance data into a usable – and actionable – format. I’m pretty certain that there’s more money in that than in simply reporting on the football results.