Page 6 - GIS for Science: GIS Response to COVID-19

xvi
GIS for Science
have messed up thousands of people who were connecting to it to make their own versions of the dashboard that focused on the area of interest, or they were running it to do analysis, so the dashboard became a visual hub for the public, but behind the scenes, the hosted service that was there became the tool, the disaster response and health community’s reason to make decisions. So those two were actually being driven by one set of services.
LG: That’s a good point. From the start, this whole thing was about open data and open science. So the whole service was open; people could pull the feature layers, and we also deposited all the data that went into the dashboard into GitHub, and that was available. We have to be careful about the data structure because we have all these files there that if we want to add variables or ever change the structure of these files, it affects people all over the world who are pulling those data on a daily basis and utilizing them in their own modeling tools and visualizations. And any time we change anything, it does the same thing to their scripts. So it has to be justifiable to make these structural changes. We can’t be selfish about making our own dashboard because every time we want to do something, we have to think about how it affects everybody else that’s using it.
That is such a fantastic point. We are learning so much now about culture change and science in terms of opening up our data and our methodologies and our workflows. It used to be when you were a student, you’d do something, and it wouldn’t go beyond your professor and maybe eventually a published academic paper. But today, we are getting used to the fact that thousands of people may be depending on our data and workflow, and this is such an apropos example of that.
LG: It’s true. I don’t think we do any work in my group anymore where you could publish unless all the data you use in your work can be published alongside of it. Today, we share out everything, everything’s open. I think that’s the only way to do science now. So now you actually have trust issues if you don’t make the data available. This COVID dashboard is such a great example of that. A major research interest of mine, something I was doing before this all came about, is about issues of misinformation and disinformation and the growing lack of trust in science and empirical understandings of how the world works. A lot of this is just about being a transparent source of information to the public, so that they can see what’s happening, but also for scientists to have open access to the data as well, so that they can build models that are also transparent in terms of where the data’s coming from.
Many people have been surprised to learn that this dashboard that has become the de facto “authoritative” source of information about the pandemic would come from a private university as opposed to a government agency. What about that?
LG: I think there’s value in it coming from a university. And it’s been great that our university has been so supportive and that we didn’t have to fight for the right to continue. I was conscious from day 1 of where I would accept funding and support for this to make sure it didn’t get branded with any organizations that would take trust away from it. And this was a delicate issue with the federal government because of how politicized this pandemic was from the start and the way they were censoring the CDC and some of the science. So I think that whether or not it should have been done by the government is kind of a separate question, but I do think the fact that it came out of Johns Hopkins, a highly respected institution in public health and medicine, has been a huge benefit. But I will note that it is not in the public health school or the school of medicine; it’s in the engineering school, which is also really great!
SB: So that does bring up an interesting question that you and I have talked about a few times as the number of cases grew in the US and we started to collect information.
In trying to get it down to the county level, there were a lot of challenges because all the sources were not just recording stuff differently but using different software products, and the ability to access the underlying data was a challenge. Can you talk a little about that?
LG: Yes, it’s been a huge challenge because the reporting criteria and guidelines and structure and the types of things being reported are still constantly changing, but yes, I think maybe what you’re getting at, and maybe something that’s been an interest from the start, is that we need more systematic and strategic guidelines and processes for reporting, moving forward, that counties and states can follow that all align into a system where it’s available in a timely manner. There are, of course, privacy issues, and things have to be aggregated and anonymized when they scale up. But here we were trying to collect data, and there’s some three thousand plus counties in the United States, and all the counties report data differently, the counties in a state might report differently than their state, and there are all these inconsistencies at all levels.
And then you have not only this issue of cases and deaths, but also you might get probable cases and probable deaths, and then there’s the issue of testing, and so many little things around this that were (and still are) a challenge. We’ll see a city or a county in one state that’s reporting something, and the state doesn’t even report that. And the state says something different about that county than the county says about itself. When we started this, there wasn’t a single COVID-dedicated website by anyone, by any government at the state, national, city, or county level. And now almost all of them have it, and we’re going through those and trying to go straight to those as our sources to pull data directly into this dashboard. This is, in itself, amazing and one of the remarkable outcomes of all this. I can say with some confidence that people in the world of epidemiology and health science reporting will be much better equipped for the next pandemic.
It is crystal clear that there needs to be a system in place so that when the next thing happens, not five months into it but weeks into it, everyone’s already put up their county dashboard, and here’s how we report the data, and this data is pulled from this dashboard into this centralized state dashboard, and the state dashboard data can be pulled directly into a centralized US dashboard. And everyone, every county, is reporting the same variables, at the same time period, and things like that. It’s complicated, but it can be done. We did it, you know? We have built it, and we’re just some engineering professors and students.
What should the GIS community be doing differently in the next pandemic?
LG: GIS people should have the systems in place and connected, and the data provided in some kind of consistent format, so that they can be pulled together into some centralized system that is open and public and accessible and usable. The CDC collects data from states and counties, and they don’t share it, which is useless to us. So the public can’t see it, so they don’t trust it, researchers can’t get access to it, so they can’t even use it, so I mean from day 1, the stuff we’ve been doing, even in January, people were using this to help policy makers in China understand what was going on, and it’s been used since then for every country as it got hit.
This has been a fascinating conversation. Thanks so much, Lauren, for talking with us.
LG: You’re welcome. Thanks to you and thanks to Esri for being so responsive and supportive during this time.