Covering Disruptive Technology Powering Business in The Digital Age

BIG DATA

The term is used as a buzzword by marketers and Big Brother-fearers alike—a means of peering into the mind of the internet user with clues they may not realize they’re leaving behind.

Those same breadcrumbs that hint toward terrorist activity or tell Google which products you’re most likely to buy, though, are being studied and used toward improving public health and preventing child maltreatment and sexual abuse, among other beneficial uses.

“There’s always this sense of ‘Big Brother’s watching,’ like everything I say is being interpreted by a machine or something, and there’s still an element of that, but we’re not trying to be creepy about it. We’re trying to save lives and do something helpful and expand on what others are doing,” says Christophe Giraude-Carrier, a professor of computer science at Brigham Young University.

Searching for clues

Giraude-Carrier has been sifting through social media to find hints of prescription drug abuse or signs of potential suicide. The idea for his studies came about after seeing the work that had been done tracking the spread of flu viruses by data mining social media platforms.

“There’s been some work on tracking things like the flu on social media like Twitter, but one thing we wanted to see if we could do is track issues that were different,” he says. “We wanted to focus on issues that were relevant in Utah.”

Prescription drug abuse and suicide are both significant public health concerns in the Beehive State.

To see if his studies were even feasible, Giraude-Carrier and his researchers, including some from the psychology department, scoured Twitter for two to three weeks to find tweets in English from across the country that had any form of suicidal ideation in them. The tweets ranged from the overt—“I want to end my life”—to the more subtle—“I feel alone, nobody cares,” or “The world would be a better place without me.” The number of tweets with suicidal ideation in them was compared to suicide rates from the Center for Disease Control overall and by state.

Giraude-Carrier’s hypothesis that people considering suicide would make a mention of it online was correct. The rates of social media mentions and actual rates of suicide were comparable using the CDC data.

“This suggests whatever happens online is indicative of what will happen in real life. It’s a decent sort of data,” he says.

Another study focused on prescription drug abuse, such as prescription pain medication and Adderall, an amphetamine-based drug used to treat ADHD and narcolepsy. Adderall is also frequently abused by students who need an extra “edge” for big projects or tests—a common conception Giraude-Carrier’s research suggests is true.

“What we found is we had huge spikes of chatter about Adderall in December and the first couple weeks of April, when finals are typically held,” Giraude-Carrier says.

Research also looked at chatter about opioids, such as prescription pain medication. Social media is such a valuable source of data because it doesn’t just have content—it also shows associations between people. Researchers looked at both users and the networks to which they belonged to see if those patterns held true for the group as well as the individual.

In respect to the research on prescription drug abuse, Giraude-Carrier says he hopes their data can help bolster efforts to close down the black market of opioides and amphetamines. The state already has a database, and there’s a database through the Utah Department of Health where doctors report what of these potentially addictive drugs was prescribed and to whom.

For the research into suicide, Giraude-Carrier again cautions any use of the data would be used in conjunction with, not in competition with or as a replacement for, current prevention and help efforts. There are a number of good programs throughout the state and especially in schools, Giraude-Carrier says, but social media is unique in its ubiquitous, constant presence. People tend to be online all day, and particularly at night, he says, and suicide data suggests people struggling with thoughts of suicide also tend to be more vulnerable at night.

No intervention is currently underway using the data. An app that could be used to help pull users away from thoughts of suicide was briefly employed but abandoned out of fears of abuse—that some less benevolent-minded users could use it to push vulnerable users toward, not away from, suicide. However, Giraude-Carrier says, there are a few ideas floating around on how to use the data. For instance, school counselors or Hope Squads could be given special access to the data to help specific people within that network, he says.

The possibilities of using social media to find data and trends invisible in most social circumstances are seemingly endless, he says, both because of the way data can be tracked and the free manner in which people post things about their personal lives.

“Almost any subject we’ve considered looking into, we find information. People are just spilling their guts. That’s something we’ve talked about. Is the whole notion of privacy something that’s changed?” Giraude-Carrier says. “Is it just ignorance? Do they think they’re only talking to their friends and unbeknownst to them the whole world can, too? That part’s unclear. One nice side benefit of this is it allows us to do things like this and use [the data] in a helpful way.”

Armed with data

Researchers from outside the Beehive State are conducting big-data research using Utah’s vast technological resources. At the Qualtrics Insight Summit in February, Dr. Dyann Daley of Cook Children’s Hospital in Fort Worth, Texas, presented a study that used big data to pinpoint areas where children were more likely to be abused or neglected.

Ten percent of kids and 20 percent of babies will be victims of maltreatment, she says, and the effects from that can impact brain development and the way genes express themselves as the children grow, potentially setting them up for a lifetime of illness, poverty or abuse. By taking the addresses of places where child abuse had been reported and combining that map with maps of poverty, crime and other factors known to be commonly present in many child abuse situations, researchers were able to make a predictive risk factor analysis that showed where kids were most likely to be experiencing abuse or neglect.

Those areas were then targeted with crime prevention measures, such as increasing lighting and decreasing foliage and other things that might enshroud a house or apartment building from the outside world. Free education and training about recognizing and reporting child abuse and neglect were also given to the research-indicated areas. As of the time of the presentation, about 2,000 people had taken the training, and all reported being more willing to report child maltreatment, Daley says.

Jennifer Freyd, a professor of psychology at the University of Oregon, also presented at the summit with the results of her ongoing study into sexual assault on campus. After a spate of sexual assaults on the Oregon campus, she says, she began conducting studies to find out how prevalent sexual assault was. The result of the Qualtrics-powered survey in 2014 showed 10 percent of female college students had been sexually assaulted, and another 20 percent had experienced attempted sexual assault. Graduate students were at the greatest risk for sexual harassment, she says.

Furthermore, she says, almost 40 percent of the victims in 2014 and 44 percent in 2015 reported experiencing institutional betrayal—that is, feeling the institution failed to prevent the issue or even allowed policies, programs or behaviors that exacerbated the problem. Victims experienced higher rates of depression and anxiety, and were more likely to become academically disengaged, she says, especially if they felt that sense of institutional betrayal.

Freyd has given two presentations at the White House about campus sexual assault, and resulting programs for educating college administrators and raising awareness have spread rapidly throughout the country, including at Utah colleges. While sexual harassment and assault on campus is still a considerable problem, having the data helps to spark a conversation and bring change.

“As disturbing as this is, it is actually a moment of incredible hope that we are now able to talk about this,” she says.

This article was originally published on www.utahbusiness.com and can be viewed in full here

(0)(0)