Using AI and Data Science to Combat Health Pandemics

Anindya Ghose

By Anindya Ghose

By Anindya Ghose

As Covid 19 continues the rattle the world, scientists and health officials around the world have been puzzled by at least one issue - the exceedingly high infection and fatality rates in Italy, compared to the rest of the world. Several theories have been floated. Was it due to the disproportionately high percentage of senior citizens in Italy? But then Japan also has a very aging population, and yet the impact of Covid-19 in Japan was not anywhere near as bad as in Italy. Was it the high percentage of smokers in Italy and their lung capacity? But then countries like Greece and Russia rank higher in smoking rates and yet the extent of Covid’s penetration in these countries is relatively miniscule compared to Italy. Vodafone recently provided Italian officials with anonymized customer data to track and analyze population movements in Italy, where there is a government-mandated lockdown. A crucial insight from the analyses of telecom data was that up to 40% of residents in Milan still moved every day beyond a 300 to 500 meters range from their home, despite the lockdown. While entirely legal, the tracking had not been previously announced to residents. It was the real time data from telecom carriers that revealed the extent of non-compliance of social distancing measures. This non-compliance of social distancing and self-isolation is likely to be a significant factor in the extent of Covid-19’s diffusion in Italy. The use of such data analytics in our understanding of the causes of health pandemic is a result of significant transformation in the ability of firms to collect massive datasets and harness them using AI algorithms.

The pace at which firms have innovated to collect and process real time data is simply astonishing. Tech firms in Asia have developed mobile apps to help people check if they have taken the same flight or train as confirmed virus patients. Government officials have used such data to track down individuals who may have been exposed, screen them, and if necessary, quarantine them. Baidu used infrared and AI-powered facial recognition to screen people at airports and railway stations for fever. Machine-learning programs have analyzed social media posts and search engine query data to predict the size and speed of the outbreak in different part of China. Chinese firms are using drones and robots to perform contactless delivery and to spray disinfectants in public areas. Google’s DeepMind is using deep learning to find new information about the structure of proteins associated with Covid-19, which in turn can provide important clues to the coronavirus vaccine formula.

Thus, today’s pandemic response can be different in part due to advances in data collection and harnessing that data through AI algorithms. It is notable that a small Canadian AI startup BlueDot spotted Covid-19 nine days before the WHO alerted people to the emergence of this coronavirus. Scientists have used real-time maps and sophisticated forecasting algorithms from epidemiology to predict the number of infected people who left Wuhan and track the spread of the novel coronavirus by analyzing air traffic patterns across China.

Read the full BrandEquity article.

Anindya Ghose is Heinz Riehl Professor of Business