FUTURE TRENDS
Public agencies and private companies alike are working hard to adopt methods of data science in the interest of food safety. In 2013 the US FDA awarded a $50 million federal contract to Dynamics Research Corporation (now part of Engility, Inc.) to help move the agency into the big data era. And, given the prevalence of mobile apps and smart phones, it is not surprising that a number of efforts are mining social media data, much as Google did for influenza. The City of Chicago Department of Public Health is working with the Smart Chicago Collaborative to develop mobile applications that monitor Twitter for possible food poisoning references. The New York City Department of Health and Mental Hygiene is working with Columbia University to review restaurant-goer comments on Yelp for possible clues to a food contamination event or outbreak.
Many private sector companies are also contributing data-driven technologies and analytics to support the push to an integrated, data-driven approach to food safety. IBM recently announced a new predictive analytics technology that the company claims is capable of identifying contaminated products “within as few as 10 outbreak case reports.” Like NCFEDA, the goal of IBM’s technology is to reduce the time required to identify the likely contamination sources by days or even weeks. Predictive analytics and other algorithms look through petabytes of grocery store food sales data from retailers and distributors in search of patterns and relationships that may indicate contamination. Visualization techniques link the data to geographical information to connect suspected contaminations with clinical and lab reports, as well as other data. A pilot is being conducted with the Department of Biological Safety at the German Federal Institute. The project will process information from 1.7 billion supermarket items sold in each country.
These developments are the first steps toward the integration of the food chain within an Internet-of-Things (IoT) environment. Situational awareness of complex and lengthy food chains is currently constrained by difficulties associated with the timeliness of data collection and fusion—in fact, much data is still manually entered into systems. In an IoT environment, end- to-end data needed for both surveillance and response can be autonomously and automatically collected using the sensor-enabled network environment of the IoT. In this, hopefully, near-term future, all stakeholders in the supply chain from the farm to the consumer will have sensors and systems in place to monitor both the food as it moves through the food chain and the health data and laboratory data needed to identify and confirm foodborne illness—and most importantly the connections between them that enable the incidence of illness to be linked immediately with the offending products in the food chain and with its source.
