In the latest installment of our occasional conversations with Tarrant County newsmakers, UT-Arlington civil engineering professor Michelle Hummel explains how crowdsourced data – including cell phone location data and neighborhood reports submitted through a smartphone app – could help historically underserved communities plan for natural disasters.
She is testing the strategy in Gulfport, Mississippi, which has experienced flooding challenges with hurricanes and the construction of retail and housing developments. Hummel’s research project is funded by the National Science Foundation and will involve assistance from UT-Arlington professor Ramez Elmasri as well as researchers at the University of California at San Francisco and the Institute for Sustainable Economic, Educational & Environmental Design.
This conversation has been edited for length and clarity. For the unabridged version, please listen to the audio file attached to this article.
Samsel: Why don’t we start with how you got into this research and how you came to UT-Arlington?
Hummel: I started my research as a graduate student in the San Francisco Bay Area, and through that project, I was looking at the impacts of sea level rise on coastal communities. I studied how different types of critical infrastructure, like wastewater treatment facilities and transportation networks, would be affected due to sea level rise and the flooding that it caused, and how that would affect not only the people who are living along the coast and might be experiencing flooding themselves, but also how that would kind of cascade through these infrastructure networks and impact others, even people who lived further inland but who are still relying on these wastewater facilities or these roadways in their everyday lives.
Following that, I came to UTA as an assistant professor and continued this sort of research, but broadening beyond just sea level rise to look at other types of flooding that also impact Texas communities, like hurricane events, river flooding and rainfall flooding as well.
Samsel: This may sound like an odd question, but what interests you about natural disasters or floods? How did you find yourself focusing on that?
Hummel: It kind of bridges the gap between the natural processes and the natural systems, and the human systems that we have in our community. These sorts of challenges come from the environmental side, where we have rising sea levels or hurricane events that are generated through these physical processes. But then they also critically impact these human systems in our communities.
It’s such a planning challenge for communities in determining, first of all, what are the threats that we’re facing? What are the uncertainties associated with that? But also, how do we respond from a political perspective, from a social perspective, a cultural perspective and an engineering perspective? Really working at the interface of those two types of processes is what interests me the most, and trying to tackle some of these challenges that I think will be major concerns for communities as we move forward.
Where are the common origin points or destination points that people are traveling between? What routes are the most critical for them? What modes of transportation are most critical?
Michelle Hummel, UT-Arlington professor
Samsel: That brings us to what you’re currently working on. How did you come up with the idea of using crowdsourced data to determine how people will be affected by a natural disaster?
Hummel: One thing that we were noticing is that a lot of communities, particularly underserved communities and low-income or marginalized communities, often feel that their concerns and priorities are not necessarily taken into account by decision-makers at the municipal level.
Oftentimes, with the existing data that we have available to us generated through satellites or gauges that different federal agencies have put into place, or through these regulatory processes that agencies like the Federal Emergency Management Agency use to map flooding, those often do not capture the very local impacts and threats that these communities are facing.
We are hoping that by using this crowdsourced data approach, we can gain additional insight from a bottom-up approach where we can really engage with the residents and document their concerns and the threats that are impacting them. (Those findings) can then help to inform future planning efforts at the county or city level, and hopefully then result in strategies that are more reflective of the community as a whole, not just the most visible parts of the community.
Samsel: For people who don’t know what it is, what do you mean by crowdsourced data? What does that entail?
Hummel: Specifically for this project, we’re using two different types of data — one I would consider crowdsourced, and the others are still crowdsourced, but “community generated” is what we call it. The crowdsourced data is generated by Streetlight InSight, which is a company that documents mobility data.
This mobility data is generated from smartphones and GPS devices. As you’re moving around, if you’ve enabled location services, it can ping these different devices and document how people are moving through the community — obviously at an aggregated level, so we’re not tracking individual people. That allows us to understand how people are moving around the city and how their mobility might be affected during a natural disaster like a flood event.
In addition to that, we’re also leveraging this community generated data by utilizing a participatory technology app that’s called Streetwyze. This allows residents to submit what are called neighborhood reports where they can actually document what is going on in their neighborhood and what resources are available to them or what resources might be lacking. They can enter text or images or videos documenting what’s going on.
For example, they could use this app to map areas that flood frequently but might not be captured through those traditional data sources. With the residents themselves generating this data, we can then integrate with the mobility data to understand how people are moving around the community, how they’re using resources and how that is changed or affected when a natural disaster occurs.
Samsel: I imagine that people don’t actually know how they move around a community because it seems like it might just be second nature to them.
Hummel: That’s the value of this: being able to track general movement patterns, without having to have people actively reporting it. I think that helps to give a better picture overall of how people are using different modes of transportation or different roads.
If we can really analyze within these neighborhoods: Where are the common origin points or destination points that people are traveling between? What routes are the most critical for them? What modes of transportation are most critical? I think that will really highlight areas where specific strategies could be implemented to make sure that those places are protected in the time of a natural disaster.
A lot of communities, particularly underserved communities and low-income or marginalized communities, often feel that their concerns and priorities are not necessarily taken into account by decision-makers at the municipal level.
Michelle Hummel, UT-Arlington professor
Samsel: Are there any specific areas or regions that you’re focusing on for this project?
Hummel: For this first part of the project, we’re partnering with a community in Gulfport, Mississippi. They’ve been experiencing frequent flooding in their neighborhoods in recent years, primarily due to new development in adjacent areas, but there’s also hurricane events that have impacted them as well.
We’re starting there and partnering with a local community group there to deploy the app for the residents and start to gather this data in that particular location. But we’re hoping that this approach can then be translated to other areas as well – other coastal cities, but also other inland areas, like perhaps Dallas-Fort Worth, to really understand how people are responding to these sorts of events.
Samsel: I know this idea stems from the idea that low-income communities and marginalized communities aren’t necessarily always included in this planning process. Does something need to happen from the city planning perspective, or does it start with community projects like this?
Hummel: Obviously, the goal of this project is trying to empower the residents themselves to collect their own data that they need and that maybe isn’t available to them. But I think the beauty of this particular app and this project is that we’re hoping to get this two-way feedback between the residents themselves and the decision-makers so that as the decision-makers see what is being generated by the community, that can inform their process. (City leaders) can also provide responses and feedback back to the community as well.
If we can open up those two way channels of communication, rather than just having a fully top-down approach, we can integrate that with the bottom-up approach and really think about: How can information and the perspectives of these underserved groups really help to advance what is going on at the city level and make it more equitable, and more effective, for reducing the overall risk of flooding and other disasters within these communities?
If we can demonstrate that this is an effective approach, and that it does have some sort of tangible benefit for the residents themselves and also at the city level in terms of how they’re planning for disasters and resilience, then hopefully that would be something that other communities can use as well, particularly those that are facing similar threats and having challenges with developing strategies that are fair across the entire population.
Haley Samsel is the environmental reporter for the Fort Worth Report. Her position is supported by a grant from the Marilyn Brachman Hoffman Foundation. Contact her by email or via Twitter. At the Fort Worth Report, news decisions are made independently of our board members and financial supporters. Read more about our editorial independence policy here.