There are far more smart sensors in the world than there are people – and we have barely gotten started with smart devices. If you take a minute to let that seep in, you’ll realize just how extraordinary this is. Factory machines, hospital beds, cars, bridges, grocery store shelves, light fixtures, farming equipment – they all are outfitted with sensors that monitor activity and generate vast amounts of data. But in most cases, after the data is collected, only a fraction of it is used. That’s too bad, because there are real stories in this data just waiting to be told.
The reason most of this data is collected but not used is that analyzing all this data is just too hard. Most people blame the data scientists, or actually the lack thereof. There is far more data than there are data scientists to analyze, explain, and help people act on the resulting insights in an easily understandable way. The Internet of Things (IoT) only promises to make this task exponentially more difficult as more devices come online. In fact, Gartner predicts that, by 2020, there will be a mind-boggling 21 billion connected things around the world.
Considering that IoT is estimated to be a $1.7 trillion market by 2020, the monetization opportunities are enormous. Companies engaged in logistics, supply chain management, manufacturing, agriculture, healthcare, and distribution are already seeing early benefits of IoT, and the longer-term potential is even more exciting.
But if something doesn’t change quickly, we are going to hit a wall. Analyzing the enormous amount of data that connected devices are yielding is already a huge challenge. To overcome these issues we need a new solution. We need data storytelling at machine scale.
Why Data Storytelling is an Effective, Scalable Solution
Data has value only if you know how to interpret it and then act on the resulting insight. Typically, that means using the data you have gathered over time to assess a given situation, anticipate what might happen next, and offer advice to either optimize or mitigate the situation.
More often than not, data is relayed in numbers, graphs, and charts – the language of data science – which can be difficult to interpret. Not to mention that the data science skills required for this type of analysis are still very hard to find and notably expensive. For end users who are just interested in knowing what happened and what action to take next, graphs and charts are not particularly helpful.
True data storytelling, enabled by advanced natural language generation (NLG), offers a better way. By automatically turning sensor data into meaningful and insightful narratives that people can read and easily understand, advanced NLG makes the required data analysis and communication possible, unlocking the true value of the IoT. Explaining this data in natural language provides any audience with actionable information that, ultimately, makes their job easier.
A large part of advanced NLG is driven by a process known as narrative analytics, an approach to analysis that is driven by specific communication goals versus the typical bottoms-up approach to data analysis. With narrative analytics, the desired narrative determines the type of analysis that needs to be performed and the data required for that analysis. The result is a natural language explanation users read and immediately understand.
How Data Storytelling Can Impact Industries
Let’s examine a few real-life IoT examples where advanced NLG can be applied to industries such as agriculture, logistics, retail, insurance, and utilities.
For farmers, rather than being forced to interpret the circles on a heat map or some other visualization, they’ll receive an email with simple, easy-to-understand natural language instructions outlining the day’s planting instructions. For logistics vendors, it would mean sending a personalized route recommendation to a driver’s mobile device so he doesn’t have to decipher a new map while on the road. Likewise, retailers could deliver personalized weekly inventory reports to individual storefront managers with shelf-stocking guidelines and other instructions tailored to the specific buying habits of that particular store’s customers.
For insurance and utility companies, there are numerous, valuable ways to leverage personal usage data to better engage and serve customers. Would you value your relationship with these companies more if they provided personalized communications based on your behavioral data? Your monthly premium statement would include insight into your driving style and (hopefully) a refund because of your safe driving habits. Likewise, your utility usage update isn’t just a bill anymore, but rather a personalized letter with advice on what temperature to set your thermostat to optimize your energy efficiency and save money. These are just a few ways that combining IoT data with advanced NLG can help companies to not only improve operational efficiency but also enhance engagement with their customer base.
IoT and the data generated from connected devices can have significant impact on how we live and work, but this will only happen if these devices can communicate with us on our terms.
Stuart Frankel is the CEO and a co-founder of Narrative Science. Narrative Science, a Chicago-based technology company, is the leader in advanced natural language generation for the enterprise. Prior to Narrative Science, Stuart was President of the Performics division of DoubleClick and was a member of DoubleClick’s senior management team when the company was sold to Google in 2008. Earlier in his career, Stuart was both a practicing attorney and CPA. Stuart earned a BS from Miami University and a JD from Vanderbilt University. In addition, Stuart was recognized as an EY Entrepreneur of the Year 2015 finalist and named to Crain’s Chicago Business 2015 Tech 50 List.
Subscribe to Data Informed for the latest information and news on big data and analytics for the enterprise, plus get instant access to more than 20 eBooks.