Contact tracing has emerged as one of the key components of Covid-19 pandemic management. In an interview with ET, Abhishek Singh, CEO of MyGov and Digital India Corporation, a government organisation leading the health ministry’s contact tracing efforts, discussed how artificial intelligence tools and data science efforts such as Aarogya Setu helped authorities. Edited excerpts:
How have AI and data science helped in Covid-19 pandemic preparedness?
Data science and AI have helped in identifying the hotspots, or where the emerging hotspots are. For instance, Aarogya Setu helps us in Bluetooth based contact tracing and location based identification of hotspots. It maps (the locations) of an infected person in the last two weeks before testing positive for Covid-19.
So, based on the number of hours these people spent at a particular location, we are able to detect hotspots and areas which are more likely to have cases in the next few days. So that’s why prediction of hotspots becomes much easier, which normally is done when you say x number of cases coming up.
With the interface developed by IIT Madras, we are able to identify a location before it turns into a hotspot, and this information is shared with the local health authorities. So, without the use of data science, these containment measures would have been difficult.
But has the app as a contact tracing tool helped?
Yes, it has. From the beginning we have been saying that our app is different from apps such as Trace-Together, or the ones used in Singapore and South Korea. We are not only taking Bluetooth contact data but also location data, and based on the location data we are gathering more information. Because of the kind of population we have, it is important to identify locations and impose restrictions and contain the spread. For that location data is important.
Can you give an example of where and how Arogya Setu helped?
Ahmedabad used this data efficiently, so did Surat. So, apart from contact tracing data, it captures the areas with the telephone tower data into granular data. Like where in a particular cell tower zone an infected person has been and the signal of all the users who were present at this location tells us the areas which are mostly likely to have Covid cases.
And this is better than the Bluetooth contact tracing because it gives the overall scenario to detect those people who might not have the app but could get infected because they were in that area.