Farm to Table with Big Data Analytics
Food is nourishment but it is also a complex adaptive system of people, practices, tools, and institutions that has been shaped since the invention of agriculture by every wave of technological change. But as much as we venerate the rich traditions around the origins, growing, selling, buying, preparing, and eating of food, there are some serious issues in the supply, production, and demand for food today. A growing global population and the diminishing set of natural resources on which they depend are forcing us to rethink the way we produce and distribute food. One promising approach to improving the quality of food and the efficiency of its production and distribution is the use of big data analytics, which has the power to solve major problems such as water conservation, food industry business issues, and may even encourage us to try new foods based on personal preferences.
“One promising approach to improving the quality of food and the efficiency of its production and distribution is the use of big data analytics”
In less than a hundred years, large-scale farming has become the norm of food production, accounting for nine percent of the U.S. economy. Producing food at scale provides a way for us to feed more people, but it is beset with challenges such as the cost of land and managing the growth and health of thousands of plants and animals at a time. But big data is giving farmers the power to collect data that leads to improvements in crop yields, equipment management, livestock health, and even how we use water.
In fact, 92 percent of fresh water used by human beings is used in agriculture, and with recent droughts hitting some of the largest food-producing states such as California, it’s vital that we’re smart about how we use the water we have. Today, through efforts between Intel and the Earth Research Institute at the University of California Santa Barbara, it’s possible to analyze the data from satellite images to gain a better understanding of how much snowpack is present in mountain ranges during the winter. This leads to greater insight into how much of that water will be available for planting in the spring. For farmers, who have often watered crops based on traditional practices, sensors in the soil and on plants provide data that give an incredibly accurate view of when plants actually need water, improving the health and yield of crops and conserving precious water in the process.
Once crops and livestock are ready to head to market, an entirely new set of challenges appear. Gone are the days of buying everything from local producers – food travels all over the world to satisfy our tastes for raspberries in March and asparagus in December. The food industry faces enormous challenges with transporting produce quickly and efficiently so that food is delivered to the consumer at the peak of freshness. The proliferation of sensors and embedded computing systems allow us to precisely measure variables like temperature in refrigerated trucks and rail cars – real-time analytics of temperature data monitored 24/7 means that if equipment or environment conditions change, distributors can make remote adjustments that preserve the quality of the food en route to our grocery stores.
Once the food is at the store, retailers can use big data analytics to improve inventory management and customer experience. Analyzing the purchasing behavior of customers has always allowed retailers to track trends that help store managers manage inventory and design more effective displays, as well as make recommendations of the types of food that pair well together. Big data analytics now helps retail stores improve shelf availability and reduce errors in item classification and placement. Intel has worked on an innovative solution to acquire images from shelves using video cameras and RFID data through low-cost antennas coupled with advanced analytics using machine learning in the cloud to track inventory more accurately and automatically. With retailers estimating their accuracy of inventory at less than 70% today, there’s clearly a lot of upside in the system.
Innovative chefs have been preparing surprising dishes for decades, so it stands to reason that all of the great food pairings have already been discovered, right? As it turns out, big data is helping us uncover new taste combinations – in fact, IBM has been working with the Institute of Culinary Education (ICE) to develop “cognitive cooking,” a complex analytic system drawing from the vast collected knowledge of chemistry, food culture and taste preferences to help chefs break new ground. This collaboration has led to the Chef Watson cookbook and food truck which debuted at SXSW 2014, serving innovative combinations such as a chocolate burrito, or apple pie with pork as the surprise ingredient.
The Future of Food
In the coming years, big data analytics will help us identify the genetic identity of plants, fish and livestock to track imports and genetically modified foods, identify illegally farmed items, and perhaps even better manage food allergies. The Barcode of Life project aims to create a genetic barcode for every species, which will help identify food and its true identity so we can better manage authenticity. Analytics can also help us create better tasting synthetic foods, such as egg-free mayonnaise and cookie dough, which will benefit those with food allergies and dietary restrictions without sacrificing taste.
Feeding the Engine of Big Data Analytics
Big data analytics is creating endless possibilities of improving the food lifecycle, but where should organizations start when designing a data strategy? I believe the answer lies in the next generation of cloud-scale applications designed to analyze the massive amount of data collected from sensors, devices and servers in the Internet of Things. By harnessing the power of big data and the intelligence of machine learning in conjunction with the agility of the cloud, organizations can gain the real-time insights they need to revolutionize the way our food is delivered from farm to table.