Stats

stats

Core poverty and inequality statistics functions.

get_stats(country='all', year='all', povline=None, popshare=None, fill_gaps=False, nowcast=False, subgroup=None, welfare_type='all', reporting_level='all', version=None, ppp_version=None, release_version=None, api_version=API_VERSION, fmt='arrow', simplify=True, server=None, dataframe_type='pandas')

Get poverty and inequality statistics from the PIP API.

This is the primary function for querying household survey-based poverty and inequality estimates. It mirrors pipr::get_stats().

Parameters:

Name	Type	Description	Default
country	str | list[str]	ISO3 country code(s) or "all".	'all'
year	str | int | list[int]	Survey year(s) or "all".	'all'
povline	float | None	Poverty line in 2017 PPP USD per day.	None
popshare	float | None	Proportion of the population below the poverty line. When set, povline is ignored.	None
fill_gaps	bool	If True, interpolate/extrapolate values for years without survey data.	False
nowcast	bool	If True, include nowcast estimates (implies fill_gaps=True).	False
subgroup	str | None	Pre-defined aggregation. Either "wb_regions" or "none". When set, routes to the pip-grp endpoint.	None
welfare_type	Literal['all', 'income', 'consumption']	Welfare concept — "all", "income", or "consumption".	'all'
reporting_level	Literal['all', 'national', 'urban', 'rural']	Geographic level — "all", "national", "urban", or "rural".	'all'
version	str | None	Data version string (see :func:~povineq.info.get_versions).	None
ppp_version	int | None	PPP base year.	None
release_version	str | None	Release date in YYYYMMDD format.	None
api_version	Literal['v1']	API version (only "v1" currently).	API_VERSION
fmt	Literal['arrow', 'json', 'csv']	Response format — "arrow" (default), "json", or "csv".	'arrow'
simplify	bool	If True (default), return a DataFrame. If False, return a :class:~povineq._response.PIPResponse wrapper.	True
server	str | None	Server target — None/"prod", "qa", or "dev".	None
dataframe_type	Literal['pandas', 'polars']	"pandas" (default) or "polars".	'pandas'

Returns:

Name	Type	Description
A	DataFrameLike | PIPResponse	class:~pandas.DataFrame when simplify is True, or a
	DataFrameLike | PIPResponse	class:~povineq._response.PIPResponse when simplify is False.

Raises:

Type	Description
PIPValidationError	If parameter values are invalid.
PIPAPIError	If the API returns a structured error response.
PIPRateLimitError	If the rate limit is exceeded after retries.
PIPConnectionError	If the network is unreachable.

Example

import povineq df = povineq.get_stats(country="AGO", year=2000) df = povineq.get_stats(country="all", year="all", fill_gaps=True) df = povineq.get_stats(country="all", subgroup="wb_regions")

Source code in src/povineq/stats.py

def get_stats(
    country: str | list[str] = "all",
    year: str | int | list[int] = "all",
    povline: float | None = None,
    popshare: float | None = None,
    fill_gaps: bool = False,
    nowcast: bool = False,
    subgroup: str | None = None,
    welfare_type: Literal["all", "income", "consumption"] = "all",
    reporting_level: Literal["all", "national", "urban", "rural"] = "all",
    version: str | None = None,
    ppp_version: int | None = None,
    release_version: str | None = None,
    api_version: Literal["v1"] = API_VERSION,
    fmt: Literal["arrow", "json", "csv"] = "arrow",
    simplify: bool = True,
    server: str | None = None,
    dataframe_type: Literal["pandas", "polars"] = "pandas",
) -> DataFrameLike | PIPResponse:
    """Get poverty and inequality statistics from the PIP API.

    This is the primary function for querying household survey-based poverty
    and inequality estimates. It mirrors ``pipr::get_stats()``.

    Args:
        country: ISO3 country code(s) or ``"all"``.
        year: Survey year(s) or ``"all"``.
        povline: Poverty line in 2017 PPP USD per day.
        popshare: Proportion of the population below the poverty line.
            When set, *povline* is ignored.
        fill_gaps: If ``True``, interpolate/extrapolate values for years
            without survey data.
        nowcast: If ``True``, include nowcast estimates (implies
            ``fill_gaps=True``).
        subgroup: Pre-defined aggregation. Either ``"wb_regions"`` or
            ``"none"``. When set, routes to the ``pip-grp`` endpoint.
        welfare_type: Welfare concept — ``"all"``, ``"income"``, or
            ``"consumption"``.
        reporting_level: Geographic level — ``"all"``, ``"national"``,
            ``"urban"``, or ``"rural"``.
        version: Data version string (see :func:`~povineq.info.get_versions`).
        ppp_version: PPP base year.
        release_version: Release date in ``YYYYMMDD`` format.
        api_version: API version (only ``"v1"`` currently).
        fmt: Response format — ``"arrow"`` (default), ``"json"``, or
            ``"csv"``.
        simplify: If ``True`` (default), return a DataFrame. If ``False``,
            return a :class:`~povineq._response.PIPResponse` wrapper.
        server: Server target — ``None``/``"prod"``, ``"qa"``, or ``"dev"``.
        dataframe_type: ``"pandas"`` (default) or ``"polars"``.

    Returns:
        A :class:`~pandas.DataFrame` when *simplify* is ``True``, or a
        :class:`~povineq._response.PIPResponse` when *simplify* is ``False``.

    Raises:
        PIPValidationError: If parameter values are invalid.
        PIPAPIError: If the API returns a structured error response.
        PIPRateLimitError: If the rate limit is exceeded after retries.
        PIPConnectionError: If the network is unreachable.

    Example:
        >>> import povineq
        >>> df = povineq.get_stats(country="AGO", year=2000)
        >>> df = povineq.get_stats(country="all", year="all", fill_gaps=True)
        >>> df = povineq.get_stats(country="all", subgroup="wb_regions")
    """
    logger.debug(
        "get_stats",
        country=country,
        year=year,
        povline=povline,
        popshare=popshare,
        fill_gaps=fill_gaps,
        nowcast=nowcast,
        subgroup=subgroup,
    )

    # Validate and apply business rules via pydantic
    params = StatsParams(
        country=country,
        year=year,
        povline=povline,
        popshare=popshare,
        fill_gaps=fill_gaps,
        nowcast=nowcast,
        subgroup=subgroup,
        welfare_type=welfare_type,
        reporting_level=reporting_level,
        version=version,
        ppp_version=ppp_version,
        release_version=release_version,
        api_version=api_version,
        format=fmt,
    )

    # Route endpoint
    if params.subgroup is not None:
        endpoint = ENDPOINT_PIP_GRP
        group_by = "wb" if params.subgroup == "wb_regions" else params.subgroup
    else:
        endpoint = ENDPOINT_PIP
        group_by = None

    # Build query params (exclude subgroup; use group_by instead)
    query = params.to_query_params()
    query.pop("subgroup", None)
    query.pop("nowcast", None)  # nowcast is not an API query param
    if group_by is not None:
        query["group_by"] = group_by

    response = build_and_execute(endpoint, query, server=server, api_version=api_version)

    out = cast(
        DataFrameLike | PIPResponse,
        parse_response(response, simplify=simplify, dataframe_type=dataframe_type),
    )

    # When fill_gaps=False (and simplify=True) filter out nowcast rows
    # pipr does this because estimate_type is only returned when fill_gaps=True
    if params.nowcast is False and simplify and not isinstance(out, PIPResponse):
        out = _filter_nowcast_rows(out, dataframe_type)

    return out

get_wb(year='all', povline=None, version=None, ppp_version=None, release_version=None, api_version=API_VERSION, fmt='json', simplify=True, server=None, dataframe_type='pandas')

Get World Bank regional and global aggregate statistics.

Shorthand for get_stats(subgroup="wb_regions"). Mirrors pipr::get_wb().

Parameters:

Name	Type	Description	Default
year	str | int | list[int]	Year(s) or "all".	'all'
povline	float | None	Poverty line in 2017 PPP USD per day.	None
version	str | None	Data version string.	None
ppp_version	int | None	PPP base year.	None
release_version	str | None	Release date in YYYYMMDD format.	None
api_version	Literal['v1']	API version.	API_VERSION
fmt	Literal['json', 'csv']	Response format — "json" (default) or "csv".	'json'
simplify	bool	If True (default), return a DataFrame.	True
server	str | None	Server target.	None
dataframe_type	Literal['pandas', 'polars']	"pandas" or "polars".	'pandas'

Returns:

Type	Description
DataFrameLike | PIPResponse	A DataFrame of WB regional/global aggregates.

Example

import povineq df = povineq.get_wb()

Source code in src/povineq/stats.py

def get_wb(
    year: str | int | list[int] = "all",
    povline: float | None = None,
    version: str | None = None,
    ppp_version: int | None = None,
    release_version: str | None = None,
    api_version: Literal["v1"] = API_VERSION,
    fmt: Literal["json", "csv"] = "json",
    simplify: bool = True,
    server: str | None = None,
    dataframe_type: Literal["pandas", "polars"] = "pandas",
) -> DataFrameLike | PIPResponse:
    """Get World Bank regional and global aggregate statistics.

    Shorthand for ``get_stats(subgroup="wb_regions")``.
    Mirrors ``pipr::get_wb()``.

    Args:
        year: Year(s) or ``"all"``.
        povline: Poverty line in 2017 PPP USD per day.
        version: Data version string.
        ppp_version: PPP base year.
        release_version: Release date in ``YYYYMMDD`` format.
        api_version: API version.
        fmt: Response format — ``"json"`` (default) or ``"csv"``.
        simplify: If ``True`` (default), return a DataFrame.
        server: Server target.
        dataframe_type: ``"pandas"`` or ``"polars"``.

    Returns:
        A DataFrame of WB regional/global aggregates.

    Example:
        >>> import povineq
        >>> df = povineq.get_wb()
    """
    params = WbParams(
        year=year,
        povline=povline,
        version=version,
        ppp_version=ppp_version,
        release_version=release_version,
        api_version=api_version,
        format=fmt,
    )
    query = params.to_query_params()
    query.pop("api_version", None)
    query["group_by"] = "wb"

    response = build_and_execute(
        ENDPOINT_PIP_GRP, query, server=server, api_version=params.api_version
    )
    return cast(
        DataFrameLike | PIPResponse,
        parse_response(response, simplify=simplify, dataframe_type=dataframe_type),
    )

get_agg(year='all', povline=None, version=None, ppp_version=None, release_version=None, aggregate=None, api_version=API_VERSION, fmt='json', simplify=True, server=None, dataframe_type='pandas')

Get custom aggregate statistics (FCV, regional, vintage, etc.).

Mirrors pipr::get_agg().

Parameters:

Name	Type	Description	Default
year	str | int | list[int]	Year(s) or "all".	'all'
povline	float | None	Poverty line in 2017 PPP USD per day.	None
version	str | None	Data version string.	None
ppp_version	int | None	PPP base year.	None
release_version	str | None	Release date in YYYYMMDD format.	None
aggregate	str | None	Aggregate name (e.g. "fcv").	None
api_version	Literal['v1']	API version.	API_VERSION
fmt	Literal['json', 'csv']	Response format — "json" (default) or "csv".	'json'
simplify	bool	If True (default), return a DataFrame.	True
server	str | None	Server target.	None
dataframe_type	Literal['pandas', 'polars']	"pandas" or "polars".	'pandas'

Returns:

Type	Description
DataFrameLike | PIPResponse	A DataFrame of custom aggregate statistics.

Example

import povineq df = povineq.get_agg(aggregate="fcv", server="qa")

Source code in src/povineq/stats.py

def get_agg(
    year: str | int | list[int] = "all",
    povline: float | None = None,
    version: str | None = None,
    ppp_version: int | None = None,
    release_version: str | None = None,
    aggregate: str | None = None,
    api_version: Literal["v1"] = API_VERSION,
    fmt: Literal["json", "csv"] = "json",
    simplify: bool = True,
    server: str | None = None,
    dataframe_type: Literal["pandas", "polars"] = "pandas",
) -> DataFrameLike | PIPResponse:
    """Get custom aggregate statistics (FCV, regional, vintage, etc.).

    Mirrors ``pipr::get_agg()``.

    Args:
        year: Year(s) or ``"all"``.
        povline: Poverty line in 2017 PPP USD per day.
        version: Data version string.
        ppp_version: PPP base year.
        release_version: Release date in ``YYYYMMDD`` format.
        aggregate: Aggregate name (e.g. ``"fcv"``).
        api_version: API version.
        fmt: Response format — ``"json"`` (default) or ``"csv"``.
        simplify: If ``True`` (default), return a DataFrame.
        server: Server target.
        dataframe_type: ``"pandas"`` or ``"polars"``.

    Returns:
        A DataFrame of custom aggregate statistics.

    Example:
        >>> import povineq
        >>> df = povineq.get_agg(aggregate="fcv", server="qa")
    """
    params = AggParams(
        year=year,
        povline=povline,
        version=version,
        ppp_version=ppp_version,
        release_version=release_version,
        aggregate=aggregate,
        api_version=api_version,
        format=fmt,
    )

    query = params.to_query_params()
    query.pop("api_version", None)

    response = build_and_execute(
        ENDPOINT_PIP_GRP, query, server=server, api_version=api_version
    )
    return cast(
        DataFrameLike | PIPResponse,
        parse_response(response, simplify=simplify, dataframe_type=dataframe_type),
    )