pudl.output.epacems

Routines that provide user-friendly access to the partitioned EPA CEMS dataset.

Module Contents

Functions

epa_crosswalk() → pandas.DataFrame

Read EPA/EIA crosswalk from EPA github repo.

year_state_filter(years: Iterable[int] = None, states: Iterable[str] = None) → List[List[Tuple[Union[str, int]]]]

Create filters to read given years and states from partitioned parquet dataset.

get_plant_states(plant_ids, pudl_out)

Determine what set of states a given set of EIA plant IDs are within.

get_plant_years(plant_ids, pudl_out)

Determine which years a given set of EIA plant IDs appear in.

epacems(states: Optional[Sequence[str]] = None, years: Optional[Sequence[int]] = None, columns: Optional[Sequence[str]] = None, epacems_path: Optional[pathlib.Path] = None) → dask.dataframe.DataFrame

Load EPA CEMS data from PUDL with optional subsetting.

Attributes

EPA_CROSSWALK_RELEASE

pudl.output.epacems.EPA_CROSSWALK_RELEASE = https://github.com/USEPA/camd-eia-crosswalk/releases/download/v0.2.1/[source]
pudl.output.epacems.epa_crosswalk() pandas.DataFrame[source]

Read EPA/EIA crosswalk from EPA github repo.

See https://github.com/USEPA/camd-eia-crosswalk for details and data dictionary

Returns

EPA/EIA crosswalk

Return type

pd.Dataframe

pudl.output.epacems.year_state_filter(years: Iterable[int] = None, states: Iterable[str] = None) List[List[Tuple[Union[str, int]]]][source]

Create filters to read given years and states from partitioned parquet dataset.

A subset of an Apache Parquet dataset can be read in more efficiently if files which don’t need to be queried are avoideed. Some datasets are partitioned based on the values of columns to make this easier. The EPA CEMS dataset which we publish is partitioned by state and report year.

However, the way the filters are specified can be unintuitive. They use DNF (disjunctive normal form) See this blog post for more details:

https://blog.datasyndrome.com/python-and-parquet-performance-e71da65269ce

This function takes a set of years, and a set of states, and returns a list of lists of tuples, appropriate for use with the read_parquet() methods of pandas and dask dataframes. The filter will include all combinations of the specified years and states. E.g. if years=(2018, 2019) and states=(“CA”, “CO”) then the filter would result in getting 2018 and 2019 data for CO, as well as 2018 and 2019 data for CA.

Parameters
  • years – 4-digit integers indicating the years of data you would like to read. By default it includes all available years.

  • states – 2-letter state abbreviations indicating what states you would like to include. By default it includes all available states.

Returns

A list of lists of tuples, suitable for use as a filter in the read_parquet() method of pandas and dask dataframes.

pudl.output.epacems.get_plant_states(plant_ids, pudl_out)[source]

Determine what set of states a given set of EIA plant IDs are within.

If you only want to select data about a particular set of power plants from the EPA CEMS data, this is useful for identifying which patitions of the Parquet dataset you will need to search.

Parameters
  • plant_ids (iterable) – A collection of integers representing valid plant_id_eia values within the PUDL DB.

  • pudl_out (pudl.output.pudltabl.PudlTabl) – A PudlTabl output object to use to access the PUDL DB.

Returns

A list containing the 2-letter state abbreviations for any state that was found in association with one or more of the plant_ids.

Return type

list

pudl.output.epacems.get_plant_years(plant_ids, pudl_out)[source]

Determine which years a given set of EIA plant IDs appear in.

If you only want to select data about a particular set of power plants from the EPA CEMS data, this is useful for identifying which patitions of the Parquet dataset you will need to search.

NOTE: the EIA-860 and EIA-923 data which are used here don’t cover as many years as the EPA CEMS, so this is probably of limited utility – you may want to simply include all years, or manually specify the years of interest instead.

Parameters
  • plant_ids (iterable) – A collection of integers representing valid plant_id_eia values within the PUDL DB.

  • pudl_out (pudl.output.pudltabl.PudlTabl) – A PudlTabl output object to use to access the PUDL DB.

Returns

A list containing the 4-digit integer years found in association with one or more of the plant_ids.

Return type

list

pudl.output.epacems.epacems(states: Optional[Sequence[str]] = None, years: Optional[Sequence[int]] = None, columns: Optional[Sequence[str]] = None, epacems_path: Optional[pathlib.Path] = None) dask.dataframe.DataFrame[source]

Load EPA CEMS data from PUDL with optional subsetting.

Parameters
  • states – subset by state abbreviation. Defaults to None (which gets all states).

  • years – subset by year. Defaults to None (which gets all years).

  • columns – subset by column. Defaults to None (which gets all columns).

  • epacems_path – path to parquet dir. By default it automatically loads the path from pudl.workspace

Returns

The requested epacems data