Taking a composite of satellite images, latest census, and cell phone data, and using machine learning we now know exactly what the population landscape looks like. Its highly complex. My home town, Lahore, seen as the large red color, is just 12 million alone 2/n... 
This information can be coded into distinct settlements and mapped onto censuses to get a more accurate picture of population distributions and education structures. This has already been achieved via brute force matching, statistical matching, and machine learning...3/n...
We can also map this information onto union councils (UC), the smallest distinct political unit. We also know what these UC get from the state and what the election results are to incorporate political structures as well....4/n...
We know the location of every possible public and private school. They are in tens of 1000s catering to millions of kids daily. Remote and distance learning on this scale is simply not possible but this can also be checked....5/n...
We also know the location of every health facility, starting from the lowest Basic Health Unit (BHU) to large hospitals, we can also calculate the catchment areas...6/n...
We can also infer population movements via smart phone tracking and this information is now available esp. during COVID19 phases. We also know labor movement patterns across district lines via labor force surveys...7/n...
GIVEN this information, we can theoretically model and track potential simulation spreads, but this requires massive computational infrastructure. A country like Pakistan, should have some sort of capacity to do this, given that state is thinly spread outside of urban zones..8/n
Best strategies are good to promote, but without data and understanding underlying demographic and mobility patterns, #COVID19 related progress will be mostly ad-hoc given the lockdown vs economic trade-off...9/9
@MalikAmmar @daninomics @BilalSiddiqi @DrOsamaSiddique
@MalikAmmar @daninomics @BilalSiddiqi @DrOsamaSiddique
Hi everyone, thanks for all the comments and retweets! Most of the data is not in the public domain. I have compiled it over the years for various research projects/papers. Please send me an e-mail if you want to discuss research/policy outputs 💡💡
