List new ITN questions/concerns to discuss

bertozzivill / map-itn-cube

Geospatial insecticide-treated net models for the Malaria Atlas Project

GNU General Public License v3.0

0 stars 2 forks source link

List new ITN questions/concerns to discuss #29

Closed bertozzivill closed 5 years ago

bertozzivill commented 5 years ago

[ ] in general: do we need to rake access to stock and flow? If yes, how to we then adjust use?
[x] national access in Madagascar is very different from the stock and flow mean. Does stock and flow incorporate surveys from later years?
[ ] is the definition of access meaningful?

bertozzivill commented 5 years ago

Stock and flow data prep:

-- why do # of nets not equal llins + conv itns sometimes? -- why remove conv itns from llin count after 2007? -- should we use hh_size or n_defacto_pop? --difference between n_itn_used and n_slept_under_itn? --> one is a net count, the other is a person count.

bertozzivill commented 5 years ago

stock and flow questions:

what's up with the mics3 and nodata surveys? where do those numbers come from? --Bonnie
what's with the population adjustments in "ADJUSTMENT FOR SURVEYS NOT CONDUCTED NATIONALLY BUT ON A POPULATION AT RISK BASIS" ? Some surveys not nationally representative, just a % of the population. These are # of people in the survey, not # in the country.
where do you get the IRS countries/values? can we update them? What are they used for? -- Population covered by IRS, can get from Anita or WMR

bertozzivill commented 5 years ago

Does data about net age come into play at all in this model? I think maybe no.

bertozzivill commented 5 years ago

Why convert NMCP LLINS to nets/capita if you multiply them back out again in the model?
BUG: I don't think you actually use the NMCP data values in the GP prior

bertozzivill commented 5 years ago

What's happening with the beta distributions in the NMCP/LLIN section?

bertozzivill commented 5 years ago

possible bug: think the xx[1|2|3|4] counters on quarterly net distribution may be a quarter behind.

bertozzivill commented 5 years ago

Why declare a "data" section in the model string?

bertozzivill commented 5 years ago

What should the priors be on L?

bertozzivill commented 5 years ago

Why does not inverting sigma fit so much better?