Visualizing Vancouver’s biodiversity
A prototyped data science project at UBC could help city planners envision healthy ecosystems, not just healthy neighborhoods. The challenge? Data integration and data gaps.
By Silvia Moreno-Garcia, UBC Science
Have you ever wondered what species might be living in your local park? Wondered why coyotes have suddenly appeared in your neighborhood? Or how a new development might impact local wildlife?
Tara Moreau, associate director of sustainability and community programs at UBC Botanical Garden, and Josephine Clark, regional planner with Metro Vancouver, live for these type of questions. They had worked together on the Grow Green Guide, an online web tool that supports new green thumbs in Metro Vancouver to design a gardening space and determine the best plant choices. Now they wanted to develop a different online tool. They wanted to map biodiversity in the Vancouver region to help urban planners make more ecologically sound decisions.
Moreau and Clark approached computer scientist Raymond Ng, scientific director of the Data Science Institute at UBC. Ng manages the Data Science for Social Good program, which partners students with public organizations on research projects. He thought their web tool idea might be a good fit for the program.
But neither Moreau nor Clark were sure where the students would find the necessary biodiversity data, or if there would be enough of it.
“We thought they might have to spend months hunting for data, and as part of the program they only had 14 weeks to do everything,” Clark says. “We were crossing our fingers.”
Some 10,500 unique species later, it turned out there was plenty of information to tap into. And three UBC students have built a working prototype of a web tool that could enable city planners to see what species can be found in different parts of Metro Vancouver. The tool can generate data-driven interactive plots allowing users to explore various dimensions of the data such as space, time, habitat, taxonomy level and more.
Citizen science meets data science
“Cities play a significant role in integrating biodiversity into the built environment,” says Moreau. “One in five plants are facing extinction. So the question becomes, how can we incorporate and address this extinction crisis through urban design.”
But city planners can’t do much if they don’t know what’s out there. While Metro Vancouver has access to certain biodiversity data, the web tool Moreau and Clark envision would cast a broader net and make it easier to visualize the information and make data-informed biodiversity decisions.
And that meant tapping into databases used by citizen scientists.
Citizen science has a long history. Birders jotted down the species they encountered, entomology aficionados spotted butterflies, and botany enthusiasts pressed plants long before cellphones came along. But mobile devices allow today’s nature lovers to document their findings and share them with unprecedented ease at scale. But that data isn’t always aggregated or easy to access.
The challenge facing UBC students Raghav Aggarwal (undergraduate student in computer science), Lesley Miller (data science master’s student) and Gabriel Smith (psychology PhD student) was taking all that information and mapping it on top of the Metro Vancouver region. And doing it a way that could be used by city planners.
Birds and more birds
The Global Biodiversity Information Facility (GBIF) turned out to be the motherload. The international web database houses species occurrence data from several research and citizen science projects which the students could pull into their web tool. Other data sources included academic institutions, such as UBC Herbarium, and provincial government information from tools like the BC Systems and Ecosystem Explorer.
Quickly a clearer picture of the Metro Vancouver region began to emerge.
The most frequent species recorded? Birds — though oddly enough, crows don’t show up often in the data. The species that is often ignored? Plants are under-sampled as a whole, especially non-flowering ones. And fungi are rarely documented.
“The birding community is very exuberant, so we found we had a lot of bird data,” explains Lesley Miller. “Data is what it is, you can’t change it. But it does beg the question: Is a certain species rarely found because it’s not there or because we overlooked it?”
There is also the issue of location. Sifting through the data, the students discovered species are more often logged in Vancouver than in Coquitlam, though that doesn’t necessarily mean Coquitlam has fewer species.
This is one of the first practical applications the tool could have: By allowing researchers to glimpse the gaps in the data, they could try to fill them. For example, future citizen science initiatives could prompt naturalists to pay more attention to specific species. Since fungi are severely underrepresented, one could picture a Fungi Week prompting people to tag mushrooms they find. Or naturalists could be encouraged to log species found in areas outside Vancouver.
In other words, by understanding our biases we could correct them.
The students also developed a predictive modeling feature for their web tool. Using altitude, humidity and other information the program can predict what species should appear in a certain geographical area.
“You can figure out, ’Okay, what’s supposed to be here?’ And if it’s not showing up in your data, is it under-sampled?” says Miller. “Or is it in danger?”
Planning for tomorrow
The prototype is now in the hands of Clark and her team at Metro Vancouver, who are working to refine it.
Right now, the web tool is fairly basic and it will be a while until it reaches its full potential. What the project has proven is that it’s possible to conduct these types of mapping efforts on behalf of cities. Hopefully, that means rich biodiversity information will help planners make more ecologically informed decisions and develop urban environments that conserve species.
“There’s a lot that we can learn from data science around ecosystem services,” says Moreau. “When we look at health, data science has really helped us to make integrated, informed decisions. The world of biodiversity and conservation can really be enhanced through a better understanding of data and how all of the complexity of the natural world comes together.”
About the students
Raghav Aggarwal is an undergrad in Computer Science. He is interested in exploring how data science can solve problems in his home country, India.
Lesley Miller has a bachelor’s degree from UBC’s Faculty of Land and Food Systems, where she worked in two genomics labs focused on biodiversity research. She is currently completing her Master in Data Science.
Gabriel Smith is a PhD student in UBC’s Department of Psychology, whose research focuses on the neural mechanisms of thought production.