Big Data's Ability to Tell a Story

To even suggest that the construction and communication of a story could one day be taken over by computers can lead to eye rolling and dismissive scoffs. After all, storytelling is a uniquely human activity, one that requires creativity, emotion, and a connection with the human audience. At first glance, computer could never replicate such a thing, right? The conventional wisdom, however, might be off in this case. With the rise of big data, new ways to create and tell stories have been developed, leading many to rethink what they previously held to be true about the art of storytelling. As far-fetched as it may sound, big data analytics may one day become the most prominent way to tell a story, even if we don’t realize it.

The goal of using big data in this sense is to create a platform where automated stories are generated. Most advances of this nature tend to start with the simplest of steps. Using the technology of automated storytelling, experts can put together news reports that give readers the basic facts and interesting tidbits of events as they happen. This technique is most commonly used for financial and sports stories where data can be plentifully gathered and analyzed to form a basic look at what is happening. This was most notably seen in IBM’s recent use of big data to create news reports about Wimbledon. The result was automated stories and breaking news messages that detailed the events and illustrated the statistics involved.

This approach works best with things like sports or financial reports because the events are predictable and feature structured data that can be parsed and analyzed. A regular sports report will normally convey the stats of the athletes, who won, how they won, etc. Financial stories feature similar themes with their data, with details showing how stocks are performing, the latest economic numbers, and more. Couch these stats in a generated language platform and you can produce stories that mimic those written by human reporters.

One can easily see the limits to just such a strategy. Once you get away from the basic statistics, it becomes more difficult to tell an effective and compelling story. Big data developers understand this limitation and have been working hard to make this idea work for stories involving a lot of unstructured data. That means taking videos, audio clips, social media posts, and other sources and breaking them down into something that can be analyzed and reported, which is no easy task. But if developers are able to accomplish this, it would add more variety to news stories as it moves away from the cold facts and statistics of automated reports.

Big data’s storytelling capability can go beyond simple news reports as well. Business reports are a common item in companies around the world, and they normally have to be prepared by employees, which takes time and resources. If big data were to automate this process and product business reports that convey information clearly and concisely, that would be advantageous. Some places have already had success with this strategy. Wells Fargo, for example, has used its massive collection of data to create automated business reports that give added insight into customer behaviors and how best to develop their business. Automated storytelling and reporting in this manner can be effective at tailoring the message to specific audiences. Executives will have need for different information than the marketing team, for instance. Instead of getting employees to produce two different reports, this is handled automatically. Using big data this way can also get rid of biases that might creep into a report, creating a more objective view of how things are going.

The use of big data to write stories is only in its infancy. Already we’re seeing big data platforms like Apache Spark aiding in the creation of stories and reports. Sometime in the near future, we may see this same type of technology used for writing fictional stories. It’s of course too early to tell if big data will one day replace writers completely -- after all, readers will likely have a tough time accepting a fictional account from a computer -- but the trend does indicate that it’s a possibility. As more advances are made in big data technology, we’ll get a better idea of just what kind of role it will play in storytelling.