Big Data and Nutrition in the Information Age

Julia Bird

In the information age, data is key. The amount of digital information produced is estimated to be doubling in size every two years (1). Data can come from many sources, produced by us directly and indirectly. Think of photographs that we take, or data that we record via sensors in and around the home, or for health purposes. When we shop, loyalty programs collect data about our purchases. Our browsing history is logged and what we do on social media is recorded. Insurance companies and health maintenance organizations can track information about health care usage. Any data produced in large quantities is considered to be Big Data – it is the volume that is important. If we can tap the flow of information, the wide availability of Big Data can unlock possibilities for us to use it to improve nutrition. 

The website “Google Trends” offers a way to look at public interest in a variety of different topics including nutrition. It is based on Big Data from search phrases used in the world’s largest search engine, www.google.com.  This tool can be used to time nutritional and behavioral interventions to match when people are looking for particular information about nutrition. For example, there is a certain pattern found when people search for “weight loss” – the greatest interest is during the first weeks of the year, coinciding with the time that people make New Year’s Resolutions, and it drops off to its lowest point in December. Or searching for different fad diets, one can view the rise and fall of the Atkins, Paleo and Whole30 diets over the past decade. This form of Big Data, coupled with a useful search tool, provides fascinating insights into what information people are looking for.

The use of wearable and connected technologies such as step-counters, heart rate monitors and wifi-connected bathroom scales to log our daily life has made it possible for this Big Data to be analyzed to glean insights for nutrition (2). The wide availability of smartphones – actually highly sophisticated miniature computers with built-in sensors such as cameras and step-counters – has eased the way in which we collect information, both for personal use and for nutritional research. For example, dietary recall results can be improved if participants in dietary surveys use an internet-capable smart phone equipped with a camera to record food intake (3). Smartphones can be used to assess symptoms of nutrition-related diseases at multiple timepoints during the data to get a more complete picture of how they affect everyday life (4). Smartscales – bathroom scales that are connected to the internet and log body weight – can help people meet weight-loss goals, and provide a better estimate of body weight than occasional weigh-ins during clinic visits (5). And it may be possible in the future to take a photo of a meal to get an instant estimate of its nutritional content without the need to manually search through diet calculation software.

Our supermarket purchases can be revealing. There are initiatives underway to include a “traffic light” at the bottom of supermarket sales receipts to assess the healthiness of food purchases and transmit the information directly to the purchaser as a means to combat obesity. Large datasets obtained from supermarkets can also be used to detect changes in purchasing behavior or assess diet quality. In the United States, the largest federal aid program directed at low-income people is the Supplemental Nutrition Assistance Program (SNAP), and is considered a hunger safety net for 1 in 7 Americans. The program allows participants to purchase food with a specially issued debit card system. In a recent study, the purchasing history of almost 100,000 households was analyzed to investigate dietary quality of SNAP participants and non-participants (6). Household purchasing history is a rich source of Big Data. The results showed that American households “show room for improvement” regarding their dietary choices, and that SNAP participants tended to have a higher intake of soft drinks and sodium than people not participating in SNAP (6).  

A recent intervention investigated whether introducing a tax on soft drinks would affect the price of soft drinks and how often they were purchased{Silver, 2017 #6}. By analyzing beverages purchased in more than 15 million transactions in Berkeley, California, researchers were able to discover that most stores passed the soft drink tax onto consumers by increasing their price, and this may have caused a 20 percent decrease in soft drink consumption after the tax was introduced (7). Likewise, supermarket purchase data was able to determine that changes to the range of foods offered in the United States’ nutritional assistance program for Women, Infants and Children (WIC) had a positive effect on participants’ diets (8){Andreyeva, 2016 #7}{Andreyeva, 2016 #7}.

The use of Big Data shows so much potential to help us find out more about which diets and dietary components are good for us, and how we can make changes to improve our diet and health through nutrition. However, organizations collecting this data need to tread wisely in terms of how the data is used and who has access to it. Small changes in behavior, such purchasing folic acid supplements to prepare for a pregnancy or researching a certain disease online, signal changes in health status that should be kept confidential. Consumers should be aware of the terms of use of data about them when they subscribe to any new program or service. The risks of accidental disclosure of personal data that has been collected should be weighed against the benefits of the program or application using it. Despite these risks, the strength of Big Data is that it is collected unobtrusively, and provides a more nuanced and complete digital view of our lives than has been available in the past.  

1. IDC, The Digital Universe of Opportunities: Rich Data and the Increasing Value of the Internet of Things. 2014.
2. Almalki, M., K. Gray, and F.M. Sanchez, The use of self-quantification systems for personal health information: big data management activities and prospects. Health Inf Sci Syst, 2015. 3(Suppl 1 HISA Big Data in Biomedicine and Healthcare 2013 Con): p. S1.
3. Hongu, N., et al., Usability of a smartphone food picture app for assisting 24-hour dietary recall: a pilot study. Nutr Res Pract, 2015. 9(2): p. 207-12.
4. Zia, J., et al., Feasibility and Usability Pilot Study of a Novel Irritable Bowel Syndrome Food and Gastrointestinal Symptom Journal Smartphone App. Clin Transl Gastroenterol, 2016. 7: p. e147.
5. Steinberg, D.M., et al., The efficacy of a daily self-weighing weight loss intervention using smart scales and e-mail. Obesity (Silver Spring), 2013. 21(9): p. 1789-97.
6. Grummon, A.H. and L.S. Taillie, Nutritional profile of Supplemental Nutrition Assistance Program household food and beverage purchases. Am J Clin Nutr, 2017.
7. Silver, L.D., et al., Changes in prices, sales, consumer spending, and beverage consumption one year after a tax on sugar-sweetened beverages in Berkeley, California, US: A before-and-after study. PLoS Med, 2017. 14(4): p. e1002283.
8. Andreyeva, T. and A.S. Tripp, The healthfulness of food and beverage purchases after the federal food package revisions: The case of two New England states. Prev Med, 2016. 91: p. 204-210.