Notes from UF

[jpfairbanks]

Design Decisions

1. We decided on a contact matrix by thinking about the definition of "multigenerational family" It means that parents interact with both their children and parents and that the children interact with their grandparents. Then we confirmed with Adarsh that Indian culture has more significant multi-generational families.
2. We discussed normalization heavily. The conclusion is that because Beta in the original disease dynamics is a composite of the "fraction of pairs of people that interact" and "fraction of interaction that leads to a transmission", we can't use the survey data found by TA1 directly. We need to reinterpret Beta in terms of its two factors in order to interpret the number of contacts per age strata

What went well

1. We worked closely with our embedded TA3 member from JuliaHub.
2. People were able to approach our team and get information
3. Upstreamed a lot of code that was tech debt

What went poorly

1. Development tasks took too long relative to scenario work. We could have done the scenario immediately but instead spent too much time fixing the hacks we had done previously instead of just making the particular stratification that we needed.
2. Delayed integration with other TAs.

@epatters, @olynch @p-stokes please add additional feedback.