Shrey Jain, who with fellow U of T students created the flatten.ca virus-tracking site in March, is part of a team that has developed a similar internet-based survey tool rolled out Monday in Mogadishu.
More than 400 volunteers began going door to door in the capital, asking questions not only to gauge knowledge about the virus and symptoms, but also to collect mobility data and socio-economic and demographic information about households.
The information, collected electronically on a tablet or phone, will be uploaded to a site that will go live in a few weeks and start monitoring the spread of the virus in Mogadishu. It will help the local government gauge the level of knowledge that residents have about COVID-19, as well as determine where officials should focus resources, such as the web or radio, for public health announcements.
The data will also be used to help with “a variety of other crises that humanitarian organizations are trying to support them with,” said Jain, as well as help with more detailed mapping of the city.
Somalia’s last census, in 2014, was the first extensive survey in decades, according to the country’s directorate of national statistics. About a half million displaced people are living in camps in the city.
Jain began developing the tool with his team at the beginning of April. They saw a need, “given that at the time they only had about four ventilators in the entire country, very minimal resources for testing, and had no real measure of understanding what the spread was,” he said.
Jain said his team reached out to the COVID-19 response team in Mogadishu through a doctor with swisscross.org, a humanitarian non-profit organization.
The survey is being conducted by volunteers trained by the Durable Solutions Unit, a local Somali humanitarian organization under the auspices of the Benadir Regional Administration, the local municipal government of Mogadishu. Funding for the project comes from the European Union.
When the website is up and running, people in Mogadishu will be able to self-report symptoms.
Jain thinks there could be a role for the surveys in other countries that have limited resources, or even remote Indigenous communities in Canada. The data can be collected offline and uploaded at a later date where the internet’s unavailable.
Jain was living in residence, near the end of his first year in engineering science at the University of Toronto, when he came up with the idea to create flatten.ca to gather crowdsourced data on the virus here.
When the pandemic hit, he moved back home to his parents’ house in Port Credit. The virus has also meant the end of his plan to compete for Canada at the now-cancelled International Triathlon Union world championships in Edmonton this summer.
A volunteer writes down a unique code on the door of a house in Mogadishu that will be used to track epidemiological information over time. SUPPLIED PHOTO
“We looked for a mechanism to collect data that was reasonable and feasible for us to collect,” said Jain, who also works at the Vector Institute, an independent, not-for-profit corporation dedicated to artificial-intelligence research.
“These big bodies of organizations, they do things to the best of their ability,” he said, “but the fact that they’re so big slows down the process of collecting data when in the time of COVID, collecting data urgently is a priority,” he says. “And crowdsourcing allows us to be a bit faster because it just relies on the individuals to provide it directly to the end-user.”
Fellow U of T engineering science students Arthur Allshire and Martin Staadecker are the technical leads on flatten.ca. The team has grown to include a number of public health experts and advisers including Marzyeh Ghassemi — recently awarded a Canada Research Chair in machine learning for health — and Nick Frosst who left Google Brain to run a start-up.
Montreal is an official partner with the crowdsourcing site and although Ontario’s Ministry of Health is not involved, Jain says they’re looking at ways to leverage the data collected through the platform.
There are challenges with crowdsourcing data.
Statistics Canada is using the method to collect data more quickly to respond to the social and economic impacts of the COVID-19 pandemic, according to its website. However, the response rates to voluntary surveys ranges from 50 to 60 per cent, comparing poorly to the census, which is mandatory and has a participation rate of 98 per cent, said Peter Frayne, spokesperson for Statistics Canada.
And the people participating, even if the numbers are very high, may not be statistically representative of the entire population, he said.
Although Jain acknowledges the self-reported data collected on flatten.ca is not as accurate as using test results to map hot spots, “it still gives us an idea as to what geographic regions in Canada are experiencing the effects of COVID or other repercussions,” he said, including the flu.
Correction – June 10, 2020: This article was edited from a previous version that included Nick Frosst’s former position.
Patty Winsa is a Toronto-based data reporter for the Star. Reach her via email: email@example.com