Skip to contents

Plot a Cleaned DSIR Indicator

Source: R/dsir_plot.R
plot.dsir_data.Rd

Draw an annual trend or a same-year country comparison from a dsir_data object. All checks and plotting are offline.

Usage 

# S3 method for class 'dsir_data'
plot(
  x,
  y = NULL,
  ...,
  type = c("trend", "compare"),
  country = NULL,
  year = NULL,
  group = NULL,
  facet = NULL,
  interval = FALSE
)

Arguments

x

A dsir_data object, normally returned by gho_clean(), sdg_clean() or ghe_clean().

y

Unused; must be NULL.

...

Reserved; must be empty. Customize the returned ggplot with +.

type

Required: "trend" for points and annual line segments, or "compare" for horizontal country points at one common year.

country

Optional character vector of ISO3 codes present in x. This first version supports countries, not regional aggregates.

year

Optional integer-valued year vector. For comparisons, exactly one year must be selected or already present in x.

group

Optional column name for colour groups, such as "dim1" or "dim_sex". Trends with multiple countries default to "iso3".

facet

Optional column name for panels, such as "iso3" or "dim_age". Other population dimensions must be constant.

interval

Logical. Show reported low and high bounds as ranges? Default FALSE. No intervals are calculated, and no confidence level is assumed. Incomplete bounds are omitted with a warning; invalid bounds cause an error.

Value

A standard ggplot object using theme_dsi() or theme_dsi_facet(). No aggregation, unit conversion, interpolation or data retrieval occurs.

Details

Select one source, backend, indicator, series/measure and unit before plotting. Use normal row filtering to select other dimensions. Unmapped population dimensions must be constant even when there is only one row per country/year. Recorded aliases, such as GHE dim1 and dim_sex, may represent the same dimension. Every country/year/group/panel position must be unique; duplicate observations cause an error.

Cleaning metadata retains constant context without adding table columns. If varying dimensions or units were omitted, clean the original data again with keep_dimensions = TRUE and, for GHO/SDG, keep_metadata = TRUE. Filtering the compact table cannot recover the missing row identities. Unknown units are labelled explicitly. Metadata is evidence about the supplied data, not a certification of comparability or completeness.

Changes to plot-relevant fields marked needs_review prevent automatic plotting. Unrelated added flags or notes do not. Review derived measures and use explicit ggplot mappings for analyses outside this contract. Missing/non-finite estimates are omitted with a warning. Trends break across missing estimates and gaps greater than one year, without filling missing years. Known non-annual or truncated time periods are rejected.

Examples 

raw <- tibble::tibble(
  IndicatorCode = "EXAMPLE", IndicatorName = "Illustrative indicator",
  SpatialDim = rep(c("PHL", "LAO"), each = 3),
  TimeDim = rep(2021:2023, 2), NumericValue = c(12, 11, 10, 15, 14, 13),
  Unit = "percent", Dim1 = "SEX_BTSX", Dim1Type = "SEX"
)
data.clean <- gho_clean(raw)
plot(data.clean, type = "trend")
plot(data.clean, type = "compare", year = 2023)