Using Social Media to Monitor Water Quality and Public Health

IJC staff
IJC
October 02, 2015
Water Matters - Networking across map of North America

In most cases, people are smart enough to stay away from polluted water. They may be itching, however, to tell other people about it.

Dr. Matthew Keifer, a member of the IJC’s Health Professionals Advisory Board (HPAB), says social media platforms like Facebook, Twitter and Instagram are some of the first places people turn when they want to share and find information on public health topics.

“The best way to take the temperature of society in those areas is to look at what people are talking about,” Keifer said.

Credit: Rosaura Ochoa
Twitter Logo sculpted into the sand. Credit: Rosaura Ochoa

Social media has already been shown to be an indicator of emerging issues, especially when it comes to incidents like a harmful algal bloom in August 2014 that caused “do not drink” advisories for users of the water system in Toledo, Ohio.

This year has been no different, with ongoing concerns about blooms in Lake Erie and other water bodies --- and people sharing personal and media updates.

A tweet from July 2015 on concerns about algal blooms in Lake Erie.
A tweet from July 2015 on concerns about algal blooms in Lake Erie.

Keifer and other Board members aim to find out more about public health by identifying trends and other indicators from social media.

The Board posted a Request for Information in May, looking to tap into this field of research, “about how data from social media might be used to help characterize individuals’ and populations’ relationships with nearby lakes and streams; how they perceive changes in these dynamic water systems; and how such changes affect their health and sense of well-being.”

Keifer said local economies are affected when there’s a decline in water quality, from people not visiting the beach and eating at a restaurant to canceling hotel stays, camping reservations and fishing trips.

“We actually don’t see the direct health consequences from some of these exposures,” said Keifer, director of the National Farm Medicine Center in Marshfield, Wisconsin. “We are very likely to see social and economic consequences.”

Similar work has been done before. The U.S. Centers for Disease Control and Prevention, for instance, has found it can predict flu outbreaks by the number and location of people sharing information about the flu on Twitter.

Credit: Duncan Hull
Networking across a map of North America Credit: Duncan Hull

To date, there hasn’t been any response to the Board’s request, said Jennifer Boehme, Board secretary and an environmental scientist at the IJC’s Great Lakes Regional Office (GLRO) in Windsor, Ontario.

But a couple of non-governmental organizations have expressed interests in results of the project.

The Board is still pursuing the initiative, and developing a proposed work plan. One effort may be a study on social media conversations surrounding annual algal bloom events in Lake Erie.

“This is the sort of topic that could be expanded to include other watersheds,” Boehme said.

Keifer notes that there numerous computer programs that can crawl the Internet to gather this type of data. The challenge is to analyze the data, and figure out how it can be used to further the Board’s mandate.

From the Board’s perspective, one of the hardest questions for health researchers is tying environmental events to health outcomes. A pilot project would give a sense of whether or not social media has information on pathways.

Rather than survey questions after the fact, this is real-time information about not just the impacts itself but actions people are taking and likely to take. If we understand that, we might be able to provide better health warnings and advice to people.

IJC staff
IJC