The hardware and bandwidth for this mirror is donated by dogado GmbH, the Webhosting and Full Service-Cloud Provider. Check out our Wordpress Tutorial.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]dogado.de.
dev branch)Between 14 and 16 September 2026, IBGE placed
apisidra.ibge.gov.br behind a Cloudflare browser challenge.
The host now answers HTTP 403 to most programmatic requests, regardless
of the client: ten identical requests measured on 2026-09-18 returned
three successes and seven challenges. Every series the package fetches
was affected.
fetch_sidra_rolling_quarters() and
fetch_monthly_population() now query IBGE’s aggregated-data
API v3 (servicodados.ibge.gov.br/api/v3/agregados), which
serves the same tables and is not challenged. The legacy SIDRA paths in
get_sidra_series_metadata() are unchanged and are
translated at request time, so nothing in the public interface
moves.
sidrar is no longer a dependency. Its 0.5.1 fallback to
the v3 service rejects whole-series period selections
(p/all) and decimal modifiers (/d/), which
every path in the metadata table uses. Removing it also drops six
transitive dependencies; curl and jsonlite
take its place.apisidra (2026-09-14): every value over the 569 shared
periods is identical, including the fourteen series whose rounding came
from a /d/ modifier, which the new client reproduces
client-side.PNADCperiods.sidra_base_url overrides
the service host and PNADCperiods.sidra_timeout the
per-request timeout (default 60s).Behaviour on an unreachable API is unchanged: an informative
message() and NULL returned invisibly, never a
warning() or stop().
mensalize_sidra_series() no longer re-anchors each (Jan,
Feb, Mar)-style trio to its rolling-quarter mean as a final step. The
internal call to .apply_final_adjustment() was removed from
.mensalize_single_series() and from both halves of
.mensalize_split_series(). The algorithm now stops at
m = y0 + cum: three strictly independent month-position
sub-series accumulating from their own starting points.
Why: the legacy final adjustment used
rq_lead1/rq_lead2 to enforce
mean(m[t-2..t]) == rq[t]. Whenever IBGE published the
rolling quarter that ends in month t, the trio
(m[t-2], m[t-1], m[t]) would all shift by the same delta
rq[t] - mean(y[t-2..t]). In real PNADC series this delta is
nonzero (because microdata-derived y0 is not exactly
consistent with later-published rq values), so previously published
months kept silently moving each time IBGE released a new rolling
quarter.
Effect on output: monthly values previously published are stable across IBGE releases. The trio average no longer exactly equals the official rolling-quarter value; the maximum drift observed in 2026-04 real data was about 0.0005% relative to the published rq.
Backward compatibility: the internal helper
.apply_final_adjustment() is preserved (still unit-tested
and reachable via :::) for users who need to reproduce
historical analyses that assumed the legacy behaviour. Calling the
public mensalize_sidra_series() no longer invokes it.
message() (no warning(), no
stop()) and return NULL invisibly when the
SIDRA API is unreachable or returns an unexpected schema. Affects
fetch_monthly_population(),
fetch_sidra_rolling_quarters(), and
pnadc_apply_periods() (which falls back to returning the
data with the crosswalk applied but no calibrated weights when SIDRA is
unavailable).testthat::local_mocked_bindings)
verify the graceful-failure behaviour offline.target_totals
mocks instead of fetching live SIDRA population data.NA in the rolling-quarter input (e.g., when
IBGE has published IPCA/INPC for a month but not yet the corresponding
PNADC rolling quarter) no longer produces phantom mensalized values that
silently echoed the value of the same mesnotrim position
three months earlier. The mask now zeros out only trailing positions
after the last observed rq, leaving leading positions
intact.compute_series_starting_points(), with test coverage for
derived-series propagation through aggregates.install.packages("PNADCperiods") now that the package is on
CRAN.\dontrun{} with \donttest{} for
API examples; created runnable examples for
validate_pnadc() and
get_sidra_series_metadata()anchor parameter in
compute_starting_points_from_microdata().onLoad that globally set
data.table thread countskip_on_cran() to all API-dependent testsfst to Suggestspnadc_identify_periods(): Build crosswalk mapping
survey observations to months, fortnights, and weeks using IBGE’s
technical break rules and respondent birthdates.pnadc_apply_periods(): Apply crosswalk and calibrate
weights to SIDRA population totals with adaptive hierarchical cell
definitions.pnadc_experimental_periods(): Probabilistic assignment
and UPA aggregation strategies for improved fortnight/week determination
rates.validate_pnadc(): Input validation for required PNADC
variables.get_sidra_series_metadata(): Metadata for 86+ PNADC
rolling quarter series available from the IBGE SIDRA API.fetch_sidra_rolling_quarters(): Download rolling
quarter data from SIDRA.mensalize_sidra_series(): Convert rolling quarter
series into exact monthly estimates via matrix inversion, without
requiring microdata access.fetch_monthly_population(): Fetch monthly population
totals from SIDRA.compute_starting_points_from_microdata(): Compute y0
starting points for the mensalization algorithm from survey
microdata.These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.
Health stats visible at Monitor.