New features for releasing

[richardli]

Testing

• [ ] Comprehensive test of all algorithms and aggregation steps
  • [ ] Write up clearly what happens with missing area in the input, when it comes to aggregation
  • [ ] Check left_join deals with missing area as intended throughout.

Modeling

• [ ] Add LONO model as alternative

Plotting

• [ ] Make interval plot work with both admin 1 and 2 cases and in more general fashion

Infrastructure

• [ ] Standardize output column names, change value to mean and include a median column as well.
• [ ] Existing indicators restructure using the modulated function approach
• [ ] Finish the list of new indicators included in the first release
• [ ] New indicators to be added using the modulated function approach
  • [x] a short document on how to process a a new indicator
• [ ] Avoid loading INLA unless using spatial models. Test to see it works.
  • [ ] Suppress SUMMER warning when using direct estimates

Clean Up

• [ ] Check all examples up to date
• [ ] Check all column naming conventions are consistent
  • [ ] Replace "value" with "mean" when appropriate. Add "median" when available.
  • [ ] Replace "HT.xxx" with "direct.est", "direct.se", "logit.direct.se", "direct.lower", "logit.direct.lower", etc.
  • [ ] Replace "DistrictName" with "admin2.name.full"

[richardli]

For future

• [ ] Systematically deal with design variance being too small (@Andrea)
• [ ] Stan implementation of both FH and cluster models (with a toggle in the function argument)