Purdue professor uses social media analytics to better understand how people perceive food products, such as #milk and #eggs
Professor Nicole Olynk Widmar
Who’s talking about milk on Twitter? Eggs? What about Disney World and the USDA labs there? No, seriously, who’s talking about milk and how do they feel about it? These aren’t some nonsensical questions, but rather a serious academic inquiry by Nicole Olynk Widmar, a Purdue professor of agricultural economics.
Widmar is studying the intersection of milk (and much more) and Twitter because research around agricultural products and consumer big data is so far scant. People have long wondered, “Why do people drink milk?” Before you were able to search millions of people’s opinions via keywords, it was just assumed that the only way to reach the customer was in the grocery store – or to seek them out to complete surveys, says Widmar, who also is the associate head and graduate program chair for the Department of Agricultural Economics.
“I’m not certain that we, in agriculture, know what big data is to the rest of the world,” Widmar says. “But it’s a big process to get meaningful data; good data is not abundant nor cheap.”
Enter social media. The types of personal data users are shedding online will be the subject of Widmar’s talk, “We’re All Listening. The Multitude of Astonishing (Petrifying?) Uses of Your Social Media Data,” at Purdue’s Dawn or Doom conference, an annual event dedicated to exploring the risks and rewards of emerging technologies. Dawn or Doom ’18 will be held on Purdue’s West Lafayette campus Monday and Tuesday, Nov. 5-6. The conference, now in its fifth year, is free and open to the public.
Scraping the internet for good data, teaching algorithms how to be more like humans and figuring out who exactly prefers cage-free eggs is all in a day’s work for Widmar.
“I worry that agriculture is in a bubble,” Widmar says. “We think we know what the public thinks about X, but 1.7 million posts later and that’s not the prevailing view.”
For example, agriculture continues to debate free-range cages versus enriched cages for egg production systems. But, the clear “winner” in terms of sentiment in social media is free-range, even if the public’s understanding of free-range is debatable.
To scientists, the beauty of social media data is, for the most part, that it is “pure.” Social media users aren’t prompted by a survey question to say something – their commentary is almost organic. Researchers can see what people are saying and answer questions like what people are eating when and how they feel about it.
“I’m horrified as an individual,” Widmar says, “but intrigued as a researcher.”
The question of who is listening to what and to what end is where social media research turns to the “doom” side. In March 2018, the British political consulting firm Cambridge Analytica surreptitiously accessed private data from millions of Facebook users and utilized it to direct messages for political campaigns supported by the company, including Donald Trump’s and the Brexit vote.
This kind of thing, Widmar says, is the underbelly of social media data, but for the most part, users are aware that their #milk and other opinions are being shared publicly, even if they don’t think much about how the information they’re sharing could be used.
Artificial intelligence, too, comes into play when figuring out how to analyze big datasets. The number of tweets – or impressions – related to a particular topic can be in the millions. With AI, researchers are able to analyze more data more quickly, but there’s a catch. Researchers also have to train the AI to understand context clues, a skill humans have readily.
“Concepts like sentiment are being measured by machines, but first you have to train the machine,” Widmar says. “For example, ‘haunted’ would normally carry a negative sentiment, but when it’s in the context of Disney’s Haunted Mansion, it’s actually positive.”
Dawn or Doom ’18, part of Purdue’s 150th anniversary celebration, features four tracks: Machines, Mind, Body and Data. Featured speakers at the conference include Frank Pasquale, law professor at the University of Maryland who specializes in artificial intelligence law, Naomi Grewal, global head of insights at Pinterest, and Nicholas Carr, author of New York Times best seller “The Shallows: What the Internet Is Doing to Our Brains.” Visit the Dawn or Doom website for more information about the conference and a full list of speakers.
Writer: Kirsten Gibson, technology writer, Information Technology at Purdue, 765-494-8190, gibson33@purdue.edu
Last updated: September 18, 2018
