As societies around the world react to the COVID-19 pandemic, vulnerable communities including displaced people living in crowded camps and slum-dwellers are more at risk. Conditions like overcrowding and poor sanitation make social distancing and other measures to ward off the virus impossible hence making residents in such communities vulnerable.

Flattening the curve in camps would be extremely difficult because of the high number of people residents come into contact with daily and the already low hospital capacity. If frontline medical workers get sick, potentially hundreds or thousands of people will not be able to receive basic healthcare.

With this premise, the goal of the AI for Good Simulator is to model the behaviour of the Coronavirus in these communities such as a refugee or an IDP camp. Customised prevalent epidemiology models such as agent-based, compartmental, and network models would be used to model the communities and make predictions.

The compartmental model can incorporate our best knowledge of the transmission rate of Coronavirus taking into account the conditions peculiar to an environment. The agent-based model would not only incorporate the environmental factors as with the compartmental model but it would also simulate daily behaviours and interactions of individuals within the target communities.

The network model takes a perspective of modelling social structure and social interactions within the camp based on the inferred relationship between residents. This would allow for some insights into how altering collective behaviours and events can impact the spread of Coronavirus. The simulator team currently is obtaining information from aid workers in a refugee camp in Moria to refine these models for accuracy and better prediction.

Any stakeholder within a target community could be a user of the simulator: an aid worker managing a refugee camp, a local authority allocating resources, or even a member of the community seeking to protect themselves and their families. Using data inputs from residents and other stakeholders, the simulator maps out the human geography and interpersonal behaviour with these communities, identifying probable transmission hotspots and also predicting how the virus might spread by taking into consideration all of these environmental factors.

The major camp demographics used for this are the number of people currently living in the camp and the age structure of the people living in the camp. Other useful statistics include the gender breakdown within the camp, the ethnic groups represented, and the number of people with pre-existing medical conditions.

Beyond predicting the rate of transmission, the simulator would also predict and model how employing different interventions and strategies would affect transmission. This would then inform residents within vulnerable communities on measures best suited for their communities to protect themselves from the virus. Also, the insights from the models would be useful to advise effective intervention measures and strategic decision making to the government, NGOs, service providers, and other stakeholders.

Seeing as none of us has ever experienced a pandemic of this scale before, we acknowledge that technology is not the only solution to protecting residents in vulnerable communities from COVID-19. We, however, seek to contribute to this fight against COVID-19 by empowering the frontline workers and local community members to make the best decisions based on their knowledge.

As a team we have some experience doing humanitarian fieldwork, building ethical AI technologies, analysing complicated real-world data, writing software, modelling infectious diseases, conducting user research, and translating technical know-how to non-technical audiences.

Together, we believe we can support the residents in vulnerable communities and organisations working on the frontline to better plan their interventions, and hopefully save more lives.

All Posts
×

Almost done…

We just sent you an email. Please click the link in the email to confirm your subscription!

OK