Big data triggers predictive journalism
Biz Stone, the cofounder of Twitter, once regaled a panel audience with a story of how he read a tweet about an earthquake around 60 miles south of him — and then felt it seconds later. This simple story conveys the power of collective information and how it can give us insight into something before we realize we even need it.
With the aggregation of large volumes of data, we are getting closer and closer to predicting events even before they occur. While data journalism (numerical data used to produce and support information) is already being widely used, and computational journalism (using robots and algorithms to aid in the creation of news stories) is becoming increasingly common, my prediction is that predictive journalism is next in line. By using available data, journalists will be able to orchestrate predictions and write tomorrow’s headlines and stories accordingly. In addition, data visualizations will likely play a bigger role in predictive news than historical news, as journalists will be able to display why they are forecasting certain events.
Let’s examine some of the common news sections and how big data might (or might not) allow journalists to predict breaking news — even before it happens.
Financial news: Can we predict the next big financial crash? The financial industry is notoriously murky, but with the help of big data, companies are now able to have far more reference points on risky clientele or patterns of fraudulent activity on a micro level. On a macro level, big data allows financial companies to continuously monitor vast amounts of data and detect patterns in the market, giving journalists a clear signal if a downturn is around the corner.
World news: From civil wars to climate change, do we have enough insight into human behavior to know how our collective actions will affect tomorrow? While human thought is not yet trackable, human behavior is becoming more predictable with the use of algorithms. For example, two MIT students recently devised an algorithm they call the Data Science Machine that is able to approximate “human intuition.” It can construct models to predict human behavior in 2 to 12 hours versus what usually takes humans several months. This is the type of technology that will be able to help journalists understand what influences human behavior and make predictions on where there might be another violent outbreak or possible energy crisis.
Medical news: The medical industry is routinely dealing with several hurdles when it comes to publicly available data due to fragmented data sources and regulatory requirements. However, by using aggregate search data (similar to ARGO, a flu tracker tool), these types of data sets can assist journalists on making headline predictions of whether or not a city will be hit by an outbreak of an infectious disease or virus.
Weather news: Watson, IBM’s cognitive computing machine, is hard at work analyzing large volumes of weather-related data points from the recently acquired Weather.com. While this information will be publicly available to anyone, journalists can leverage this data to make predictions on regional events or perhaps travel delays or even traffic congestions days ahead of time.
Entertainment news: Will Selena Gomez and Justin Bieber get back together? Well, there are some things data just can’t help us predict.
I’m excited to see today’s data being used for tomorrow’s headlines in 2016.