Releases: WorldHealthOrganization/anthro
Releases · WorldHealthOrganization/anthro
v1.0.1
v1.0.0
General
- Improved prevalence computation code, it now consumes slightly less time and is better organized. In addition test-coverage and documentation is improved.
- Cluster/strata/sw values in
anthro_prevalence
now support an explicitNULL
value to better control for (missing) cluster/strata/sw information. - New API functions are exported, mainly useful to use complicated portions of the logic in other packages, not meant for end-users.
nest = TRUE
is now set for the survey designs inanthro_prevalence
. See the documentation ofsurvey::svydesign
for more information.- Removed
headc
,armc
,triskin
andsubskin
from parameters ofanthro_prevalence
as they were not used for the final output. _pop/unwpop
values in prevalence calculation are now 0 instead of NA if no values are present for that group.
Bugfix
- Previously, if one level in a group only has one observation, all
stdev
columns inanthro_prevalence
areNA
. Now only the levels with one observations yieldNA
, for the others thestdev
is still computed.
v0.9.4
v0.9.4
v0.9.3
General
- Internal code improvements.
Bugfix
- Z-scores are now only computed for
age < 60 months
instead ofage <= 60 months
. - Z-scores for wfl were previously also computed when
age >= 60 months
and<= 1856 days
.
v0.9.2
- Age in days is now rounded half to even (e.g. 730.5 days = 731) before joining
the data with the reference values. This is in line with previous
implementations and in particularly relevant for data points with
age exactly 24 months. Previously the example above was converted to 730 days and with
this release it is converted to 731 days (#17). - The cleaned
measure
andsex
variables are now part of the output
data.frame ofanthro_zscores
(#20, #24). - Removed
covr
from suggests dependencies. - Fixed a typo in the docs (#15).
v0.9.1
This is the initial release of the anthro package. Feedback is very welcome!