Environmental Justice in New Jersey – An Exploration of Twitter as an Early Injustice Detection Tool
Presentation Type
Poster
Faculty Advisor
Danlin Yu
Access Type
Event
Start Date
26-4-2023 11:00 AM
End Date
26-4-2023 12:00 PM
Description
In recent years, mounting awareness of the discriminatory distribution of environmental hazards and ecosystem services has increasingly placed environmental justice at the forefront of discussions on sustainable development. However, current initiatives and policies revolving around the subject are reactive in nature, responding too late to infrequently collected data from official sources and undermining the legitimacy of community science data. The emergence of big data in the form of social media and remote sensing imagery has presented an opportunity to bridge these gaps. We take the first step in examining this relationship, exploring the geographic relationship between Tweets and environmental justice factors at the Block Group and Tract level using generalized linear regressions, ordinary-least squares, and spatial autoregressive models. Using geolocated Tweets, we calculated the raw and area weighted frequency of Tweets to be used as our dependent variables. Our results suggest that there is a negative relationship between our dependent variables at socially vulnerable populations, primarily as a result of aggregation bias stemming from Census boundaries. Furthermore, controlling for spatial and aggregation bias, our models suggest a strong relationship between Twitter activity, greenery, urbanization, and air quality. In this way, remote sensing derived data provides a particularly unique opportunity to measure the relationship between environmental factors and Twitter due to the ability to generate high resolution continuous data. Going forward, we anticipate big data sources such as Twitter and remote sensing imagery may serve a dual-purpose in the realm of environmental justice – acting as an early warning system for injustices that have not yet been detected by other sources and corroborating community scientist observations that otherwise do not receive due attention.
Environmental Justice in New Jersey – An Exploration of Twitter as an Early Injustice Detection Tool
In recent years, mounting awareness of the discriminatory distribution of environmental hazards and ecosystem services has increasingly placed environmental justice at the forefront of discussions on sustainable development. However, current initiatives and policies revolving around the subject are reactive in nature, responding too late to infrequently collected data from official sources and undermining the legitimacy of community science data. The emergence of big data in the form of social media and remote sensing imagery has presented an opportunity to bridge these gaps. We take the first step in examining this relationship, exploring the geographic relationship between Tweets and environmental justice factors at the Block Group and Tract level using generalized linear regressions, ordinary-least squares, and spatial autoregressive models. Using geolocated Tweets, we calculated the raw and area weighted frequency of Tweets to be used as our dependent variables. Our results suggest that there is a negative relationship between our dependent variables at socially vulnerable populations, primarily as a result of aggregation bias stemming from Census boundaries. Furthermore, controlling for spatial and aggregation bias, our models suggest a strong relationship between Twitter activity, greenery, urbanization, and air quality. In this way, remote sensing derived data provides a particularly unique opportunity to measure the relationship between environmental factors and Twitter due to the ability to generate high resolution continuous data. Going forward, we anticipate big data sources such as Twitter and remote sensing imagery may serve a dual-purpose in the realm of environmental justice – acting as an early warning system for injustices that have not yet been detected by other sources and corroborating community scientist observations that otherwise do not receive due attention.