Now that the virus is in India, a thread 👇🏼 on the potential geography of the virus spread based on known migration patterns. No need to panic but some areas need to be even more prepared than others in terms of medical capacity. First map below shows known hotspots. 
This map on international migration is important to know first contact, which is pertinent for Punjab (potential Italy source) and Kerala. Of course, known transmission cases also happened from students and tourists, unconnected with standard migration routes.
So there was a sudden lock-down announced and there was chaos for migrant workers. This table below helps understand why the crisis was felt more in Delhi, Mumbai and Surat- all cities with large share of inter-state migrants and most importantly migrants with RURAL connections.
But some migrants still managed to scramble back home. So which districts should be on high alert for potential virus spread? Unfortunately, there is only old data from the 1990s but the only one available (yes, Census 2011 has still not released this!). Let's assume persistence.
Since Mumbai and Delhi are key centres in recent reverse migration scenario, lets look at them. That's a lot of districts but map shows you clear concentrations which need more attention.
Less worry about transmission from Kolkata to other states but more so for Surat. Ganjam in coastal Odisha better be on high alert (past instance includes AIDS transmission via Surat).
Less worry of inter-state rural transmission from Bangalore, Hyderabad, Ahmedabad, Pune, though Ahmedabad has an important source belt in Rajasthan with many reports of people walking back home.
Important source region clusters need to have strong communication about how to deal with the crisis and return migrants, in a way that does NOT lead to stigmatization of potential virus-affected migrants.
To repeat, this is only to map out potential scenarios given the circumstances. Influenza 1918 spread in interesting ways out of Bombay and weakened by the time it reached Calcutta as seen below. Below is Fig 5. of Chandra and Noor (2014) paper: ncbi.nlm.nih.gov/pmc/articles/P…
