September 21, 2017

Using social media data to help measure smoke exposure

By: Moira Burke, Facebook research scientist, Bonne Ford, Colorado State University research scientist, Jeffrey Pierce, Colorado State University professor

When natural disasters strike, people turn to Facebook to understand what’s happening and to share important safety information. The same applies to wildfires, where people often discuss the locations of fires and where they see smoke and haze. Nearly 50,000 wildfires have already burned this year.

Particles emitted from wildfires can travel hundreds of miles and cause a variety of adverse health effects, including respiratory irritation, asthma and heart disease. That’s why public health organizations need accurate data about smoke exposure to understand who was exposed and to measure safety risks.

However, it’s difficult to pinpoint how far particulate matter from fires travels. Ground-based monitors are sparse in parts of the U.S., and satellites can’t always tell the altitude of the smoke particles.

Posts on Facebook can help signal how far the particles from a fire have spread and can help researchers and health organizations understand the effects of smoke exposure on a population.

Working with researchers from Colorado State University and the National Center for Atmospheric Research, we found that incorporating aggregated and de-identified Facebook posts into traditional analyses improved our ability to track changes in air quality stemming from dense smoke plumes and large wildfires. Using Facebook data led to better results assessing smoke concentrations and identifying the location of potential exposure than traditional methods like satellite observations and chemical transport modeling.

In the study, we counted the percentage of people in a city who used terms like “haze” or “smoke” (but not “cigarette” or “barbecue”) each day during the summer of 2015, and then compared those de-identified data to traditional air quality measurements from satellites and ground sensors.

We found that in the majority of cities in the U.S., the Facebook data improved the models of smoke exposure and can help in places with sparse ground-sensor data.

One surprise was that people were more likely to notice smoke on sunny days than cloudy days. This means that people might be more at risk to health hazards on cloudy days, if they do outdoor activities without realizing the air quality is poor. Traditional methods of estimating smoke concentration don’t tell you whether people notice air quality problems; the Facebook data were a good proxy for how much people actually perceived the smoke, and it showed us that people might not be as prepared for health hazards on cloudy days.

This animation illustrates Facebook users posting about wildfire smoke (indicated by red, yellow and orange). These posts are compared with standard satellite measurements of smoke in the atmosphere, and surface measurements of particles (colored circles), during August 2015. Credit: Bonne Ford

 

You can read more about the research in this study published in the journal Atmospheric Chemistry and Physics.

We believe that Facebook can be a valuable source of information to help researchers solve problems more efficiently and effectively.