Package index • polished

Download POLIS data

The download entry point and the table catalogue it fetches: a one-call pull that works around POLIS’s pagination and date-filter quirks, with on-disk caching, resume, and post-download completeness verification.

get_polis_data(): Download POLIS tables
polis_tables_mapping: POLIS table catalogue

Configuration & pipeline

The shared settings object every cleaner takes, and the orchestrators that run the whole cleaning set in one call (in memory, or from a directory of raw files).

polis_config(): Build a POLIS pipeline configuration
polis_active_config(): The session-active POLIS configuration
print(<polis_config>): Print method for POLIS configuration
run_pipeline(): Run the POLIS cleaning pipeline in memory
run_pipeline_dir(): Run the cleaning pipeline from a directory of raw files

Clean the surveillance streams

One cleaner per POLIS stream. Each standardises names, sanitises dates, derives the analytic variables, cleans geography and dedups to one row per POLIS id. clean_virus() is the exception: it builds the positives dataset from the already-cleaned AFP and ES outputs.

clean_afp(): Clean POLIS AFP case data
clean_human_spec(): Clean POLIS human specimen (laboratory) data
clean_es(): Clean POLIS environmental surveillance data
clean_sia(): Clean POLIS SIA (campaign) data
clean_virus(): Build the POLIS virus (positives) dataset from cleaned cases and ES

Population denominators

Clean the raw POLIS population reference into adm0/adm1/adm2 under-5 / under-15 / all-ages denominators, optionally reconciled against WorldPop and rolled up via boundary-validity windows. These feed the rate indicators when no population is supplied to the pipeline directly.

clean_pop(): Clean POLIS population and (optionally) reconcile against WorldPop

Virus classification

The shared decoder that turns the raw POLIS virus fields into the standard WPV / cVDPV / aVDPV / iVDPV vocabulary used across every stream.

clean_afp_classification(): Derive the fused AFP virus type and analytic classification
clean_es_classification(): Derive the AFP-style virus classification and detection flags for ES

Clinical diagnosis

Harmonise the four POLIS fields that carry the clinical cause of an AFP case into one canonical diagnosis, then derive the coarse class, the reported non-AFP flag and the follow-up outcome.

clean_afp_diagnosis(): Harmonise the AFP clinical diagnosis

Reference tables

The packaged lookups the cleaners are driven by: the raw-to-canonical column crosswalk, the data dictionary (raw or cleaned schema) it backs, the country grouping/risk reference, and the AFP diagnosis dictionaries.

polis_crosswalk(): POLIS column crosswalk
polis_dictionary(): POLIS data dictionary (raw or cleaned schema)
polis_country_lookup(): Country reference lookup shipped with the package
polis_afp_diagnosis_lookup(): AFP free-text diagnosis lookup shipped with the package
polis_afp_icd10(): AFP ICD-10 reference list shipped with the package
polis_afp_diagnosis_class(): AFP diagnosis-to-class map shipped with the package

Recover & reconcile geography

Fill missing admin names and GUIDs without overwriting present values or fabricating on ambiguity (from EPIDs, from a district shape, or from coordinates), plus the admin-name fixers.

impute_geo_from_epid(): Recover administrative geography from the EPID
reconcile_admin_guids(): Reconcile case admin names and GUIDs against the long district shape
impute_geo_from_coords(): Recover missing admin from coordinates, in place
impute_missing_coords(): Place a random point inside the district polygon for cases missing coordinates
fix_geo_names(): Normalise admin names on a cleaned data frame
polis_fix_geo_names(): Apply geographic name fixes to a character vector
polis_geo_name_fixes(): Geographic name-fix lookup table

Clean administrative spatial data

Read WHO ADM0/ADM1/ADM2 boundary layers from any format, standardise the names and repair the geometry, then write cleaned shapes plus a year-expanded long shape, with point-to-admin recovery.

process_spatial(): Clean WHO administrative spatial data
create_long_shape(): Expand admin shapes to one row per active year
get_admin_info_from_coords(): Recover administrative info for point data via a spatial join

EPID building blocks

The exported pieces the EPID-geography cascade is assembled from: parsers, the sibling-record lookups, and the country resolver. You rarely call these directly.

epid_split(): Split an EPID into its component segments
epid_country_code(): Extract the country code from an EPID
epid_prefix(): Geographic prefix used for prefix-matching
epid_strip_contact(): Separate a contact EPID from its base case EPID
build_admin_ref(): Build an EPID -> admin-value reference (most-recent-per-EPID)
build_prefix_ref(): Build a (prefix, year) -> unique admin-value reference
resolve_epid_country(): Resolve an EPID country code to a country name

Records, dedup & types

The shared primitives the cleaners compose from: column standardisation and ordering, keep-latest dedup, the business-key tripwire, synonym remapping, full-pull reconcile, and column type inference.

standardise_names(): Standardise POLIS column names
order_columns(): Order columns: identifiers, then location, then time, then everything else
polis_upsert(): Upsert by Id, keeping the latest record
collapse_business_key(): Collapse business-key duplicates, keeping the latest record
flag_ambiguous(): Flag (do not drop) rows whose business key spans multiple Ids
remap_synonyms(): Remap merged EPIDs to their canonical value
reconcile(): Prune records absent from a full pull (reconcile)
auto_parse_types(): Infer column types after cleaning, then optionally layer factor detection
detect_factors(): Detect factor-like character columns (low-cardinality only)

Indicators

Compute the WHO POLIS surveillance indicator catalogue (NPAFP rate, stool adequacy, timeliness, dose history, environmental, virus, SIA and composite families) from cleaned case / ES / virus / SIA / lab tables. Call available_indicators() to browse the catalogue, optionally by family.

calc_polio_indicators(): Calculate polio surveillance indicators (the POLIS indicator catalogue)
available_indicators(): Dictionary of available polio surveillance indicators

Data-quality checks

Per-dataset checks that flag duplicates, blank keys, unreconciled GUIDs, out-of-range values and date-ordering problems straight from the cleaned tables, exported as a styled Excel workbook (one tab per check), plus the ES-specific quality helpers. checks_pop() documents the POLIS-vs-WorldPop population reconciliation instead.

checks_afp(): Run AFP data-quality checks
checks_es(): Run environmental-surveillance data-quality checks
checks_hum_spec(): Run human-specimen data-quality checks
checks_sia(): Run SIA data-quality checks
checks_virus(): Run virus/positives data-quality checks
checks_pop(): Run POLIS population data-quality checks
write_checks_excel(): Write a checks result to an Excel workbook
es_missingness(): Summarise missingness in key ES surveillance variables
validate_es_sites(): Flag ES site names absent from a reference site list

Other surveillance processing

Independent-monitoring (IM), LQAS campaign-monitoring, and SIA round-quality processors.

process_im(): Process raw Independent Monitoring (IM) data into missed-children rates
process_lqas(): Process raw LQAS lots into classifications and district pass rates
process_sia_quality(): Process the POLIS SIA campaign-quality tables (LQAS + IM)

Project workspace

Scaffold a project on disk — either the full domain-numbered pipeline (init_polis_pipeline(), with a wired .Rprofile and download/process scripts) or the lighter generic zones (init_polis_project()) — stream the cleaning pipeline into it, and read partitioned slices back out.

init_polis_pipeline(): Scaffold a full polished pipeline project
init_polis_project(): Create (or re-open) a data project
load_polished(): Read a country / period slice of the polished outputs
project_path(): Build a path inside a project zone
clear_cache(): Clear a project's regenerable cache
print(<polis_project>): Print a polis_project