Nearly a decade ago, Brian Solis declared, “Social media is about sociology not technology.” Five years later he updated that to say, “Social media is about social science not technology.” Last week the findings of ground breaking research were presented at the ACM Conference on Computer-Supported Cooperative Work and Social Computing in San Francisco. Georgia Tech researchers, led by Munmun De Choudhury, analyzed social media data via Instagram posts to develop a predictive model for identifying “food deserts,” areas characterized by lack of access to healthy food options. In the past, their identification has been dependent on surveys and self-reported; even anecdotal evidence from limited sample sizes. This team’s model delivers results with high accuracy (>80%).
This research allows for new areas of social media analysis for common good. It admittedly has its limits, but it also offers new access to information that was previously hidden. NPD group has been tracking the way Americans eat for decades, recruiting individuals and families to keep journals of what they eat to extrapolate for restaurants and food manufacturers to identify trends and changes in eating behavior. It was years ago that I heard Harry Balzer identify the rise in bar consumption as a snack and even a meal substitute. The Georgia Tech research widens the net, with another aim in mind.
From the research, they learned of regional differences in the most popular food posted, but we also learn that fruits and vegetables are part of nearly half of posts from people in non-food deserts, while that drops to one third in food deserts. Another interesting finding is that while total calories did not vary significantly between the two groups, the amount of fat, cholesterol and sugars was 5 to 17 percent higher in food deserts.
This information can not only help shape public policy, but also drive business decisions for supermarkets that see it as an opportunity. As we consider its relationship to the obesity epidemic, further research can look at physical activity and identify “sloth zones” of inactivity.
Using social media to address obesity and food health has been a growing area. A year ago, Bolthouse Farms launched its #urwhatupost campaign to get fruits and vegetables tagged more as #foodporn. Recently a study looked at how emojis can play a role in encouraging healthier choices by children. This is just the beginning, and with regular meetups like #RDchat and #FoodFri I see it continuing to grow.
Earlier I wrote about how political campaigns are mining social data like never before. We are just starting to scrape the surface of what is possible. For marketers, social media analytics are well past followers and shares. Engagement will continue to be important, but there are new ways to understand and identify characteristics that are consistent with your advocates. This will allow for better content, targeting and engagement that will lead toward greater marketing success.