NAM 2011-15

[QLLZ]

Data pull: HASH: 6772ea9da4f43ddc3da11c494a573af2ca3dd3c8 config

[QLLZ]

log file

[QLLZ]

rerun from data pull: data pull on dev: HASH c9a89d4ac5e4f63cb75fa5c006a900ece2839199

log file

[QLLZ]

preprocessed_data file was created manually by running the set_parameter R script to skip the error in get_country_admin_units.

HASH: c9a89d4ac5e4f63cb75fa5c006a900ece2839199

[QLLZ]

Model running: HASH 6772ea9da4f43ddc3da11c494a573af2ca3dd3c8

[QLLZ]

Country data report

[javierps]

Convergence and Rhats ok. Little sub-national data. Case and rate maps look a little pixelated. The 0 population grid dropping is causing wholes in the grid in the south western border.

Check discrepancy between the disaggregated shapefiles and the gam input figures which suggest different spatial coverage of sub-national data.

Suggestion: Second opinion. I would lean in favour of no-w run given the large size of the country and little sub-national data/

[eclee25]

Convergence, rhats, and observation fits look okay but there is discrepancy in GAM input and rate maps.

Opinion: Temp Approve and rerun with spatial_effect: 'no' model

[QLLZ]

Config

Data pull:

HASH: 211913e0eb7d570ed62c48bc1ae9ecd6671a6cbc

[QLLZ]

Failed stan model run log file

[QLLZ]

Rerun on dev_u_combs_fix

HASH: e56580fa5ddab00293e31ff90139351f80ae3f6c

[QLLZ]

country data report

[eclee25]

Instability in sd_w traceplot but all other diags look good. Substantial case counts are only present in one year. Approve

[eclee25]

Minimal subnational data but high sd_w -- try no-mixture sd_w prior

[QLLZ]

Data pull, model run:

HASH: 05f598b95b5d42ac0049963f8a0b7e04d6f1dbd7

[javierps]

Some issues with rho, but std_w and ws look OK. All Rhats OK, although fig. 26 shows some disagreement between chains, but variance is probably high so that Rhats pass. Given the little amount of data don't think that we can do much better.

Suggestion: accept.

[eclee25]

Approve