Pittsburgh Plans For Air Quality Threats With Predictive Analytics


The city of Pittsburg is adding to its resiliency plans with a predictive analytics initiative focused on air quality.

Pittsburg’s resiliency plan is known as ONEPGH and is supported through a partnership with the group 100 Resilient Cities, a project of the Rockefeller Foundation. Resiliency plans are designed to help cities prepare for a variety of issues including climate change, demographic change, and public safety concerns. The plans are designed to give city officials a playbook that will aid decision making in the event of a significant threat or change to the city so that officials can limit the impact on residents and maintain public services.

On Monday, the city along with representatives from the University of Pittsburgh Graduate School of Public Health and Intermedix hosted a preparedness workshop that looked at how emergency response technology could work with predictive analytics to help prepare the region for something similar to the killer Donora smog event from 1948. During that time, smog combined with a heat wave to compound the threat to vulnerable individuals and caused a number of deaths.

The workshop, facilitated in partnership with 100RC, hosted more than 50 public health, emergency management and air quality professionals from Western Pennsylvania.

The University of Pittsburg has an open source data modeling system called FRED (A Framework for Reconstructing Epidemiological Dynamics) that can recreate major events like what happened in Donora virtually, so that scientists and first responders can study how to improve emergency response plans. FRED was initially created to look at how infectious diseases spread throughout populations, but has since been expanded to look at non-infectious diseases, as well as social and environmental factors that affect health.

In the workshop, researchers demonstrated how they use FRED to look at the clinical impact of heat and smog on city residents. “FRED allows us to pinpoint critical conditions and the effect of potential interventions to better educate response efforts. For example, we can use the model to predict how many instances of acute respiratory disease warranting a 911 call would occur in the context of this environmental event,” explains Mark Roberts, M.D., M.P.P., Chair of the Department of Health Policy and Management and director of the Public Health Dynamics Laboratory at Pitt Public Health.

Researchers are using the data from FRED and feeding it into Optima Predict, a predictive response planning and simulation software created by Intermedix, to determine the appropriate resource allocation and deployments of emergency response personnel.

The workshop also explored communication and coordination challenges between first responders, emergency management and critical infrastructure organizations, such as hospitals and utilities in order to find ways to address service gaps.

The issues discussed at the workshop will inform future adjustments to the city’s resiliency strategy which was released in March. “Using technology and professional experience together is a great way for us to model how systems interact, and we can also use different scenarios and circumstances to replicate the process in the future,” said Rebecca Kiernan, Senior Resilience Coordinator for the City of Pittsburgh.