Where Narrative Science scores is its ability to scan and analyse a huge amount of data, big data as it is called, in a comparatively short span of time. Journalists have little time or energy to sift through such massive amounts of facts and figures. Moreover, Narrative Science can search the data and pull out significant spikes or trends and weave a story around it. It can, for instance, scan stock data and pinpoint a company that is doing remarkably well, a company not in the public eye and, therefore, easy to miss by a human researcher. This is, therefore, applicable to other areas where statistics and number crunching are essential, sports roundups, for instance, consumer data and sales figures, which show a definite trend, or even medical research.
All this has raised legitimate concerns that Narrative Science (and its imitators in other countries) with the ability to generate cheap, almost instant content, will make human writers obsolete. There may be some truth to that, but the answer is actually quite complex. We all know data is valuable, but what Narrative Science does is focus on the most valuable partthe insights that the data offers, the stories that emerge from their data mining. However, the software used can only highlight certain aspects or trends, it needs a human brain and journalistic experience to make it a readable piece of prose. In other words, what the company sells to clients is produced by cross-functional teams of computer scientists and writers. Yet, increasingly, as the software is improved and story formats established, there is the nagging question: will machines replace humans Currently, the co-founders of Narrative Science, Kris Hammond and Larry Birnbaum, professors of computer science from Yale, use the services of students and faculty at the Medill School of Journalism to create the teams of writers and coders.
Their current level of computer expertise allows them to produce a wide range of content from any data source. Where it gets tricky and demands the use of journalists and writers, is that every client wants a different version based on length, style, tone, even language. Publications, for instance, want stories that fit their house style and their readers expectations. Which requires journalists. The future, however, is where the questions arise. Once a house style is established, Narrative Science can program and customise its software to produce identical content, thus dispensing with the human factor. The advantage is that the software can absorb and analyse vastly complex information, and data sets that would boggle the human mind. It can, for instance, instantly troll the Twitter world and produce data on how millions of Twitter users feel about an upcoming election, or a sports clash between major teams or a blockbuster movie release, things that a journalist couldnt do because Twitter moves so fast, and at such a high volume. There is a reverse to this kind of massive data-mining, which is that it would be an invaluable tool for journalists.
Indeed, Narrative Science can do some astounding things with data and statistics, and even produce a story based on that, but the truth is that the best stories in the media are not data-driven and are about people and places and ideas, more specifically, descriptive narratives that capture the human experience and human emotions. Poverty can emerge from data, but not the sense of what it does to individuals, families and communities. That is where the man versus machine debate ends.