Download notebook

Obtain the economic files and their matching labels locally. Eora1/Eora26 and Eora2 are distinct source families; this guide covers the supported Eora1/Eora26 inputs.

Import#

import mario

db = mario.parse_eora(path="/path/to/Eora26_2017_bp", multi_region=True)

Result · A MARIO database. Inspect db, db.get_index("all") and db.units before selecting quantities for your analysis.

Inputs and options#

Important scope note

MARIO does not parse the full Eora MRIO release.

The supported workflows are only:

  • Eora26 as the multi-region workflow;

  • local single-region Eora tables.

Relevant source page

Automatic download is not part of the current MARIO workflow. The expected path is to work from local files that you already downloaded.

Key arguments

The key public arguments are:

  • path: one Eora file or one local dataset directory;

  • multi_region: use True for Eora26 and False for local single-region files;

  • table: mainly relevant for single-region parsing;

  • indeces: optional path to the Eora26 label files;

  • name_convention: full_name or abbreviation in the single-region workflow;

  • aggregate_trade: whether to collapse detailed import/export rows in the single-region workflow;

  • country: country selector when a directory contains multiple single-region files;

  • price: optional selector when a single-region directory contains multiple price variants.

`Eora26` versus single-region Eora

Use multi_region=True only for Eora26. This workflow expects the standard Eora26_<year>_<price>_*.txt files plus the labels_*.txt files.

Use multi_region=False for the local single-region tables. In that case MARIO resolves one file such as IO_ITA_2015_BasicPrice.txt and can usually infer whether it behaves like an IOT or a SUT.

Label files for `Eora26`

For Eora26, MARIO needs both:

  • the numeric files such as T, FD, VA, Q, and QY;

  • the label files such as labels_T.txt, labels_FD.txt, labels_VA.txt, and labels_Q.txt.

If the label files already live next to the dataset files, indeces= can be omitted. Otherwise, point indeces= to the directory that contains them. The locally held 2016 copy lacks labels_VA.txt. Its VA matrix has the same dimensions as 2017 and the other labels match byte-for-byte, so set indeces= to a complete common label folder, such as Eora26_2017_bp, when parsing that 2016 copy.

Local layout examples

There is no dedicated Eora 2 parser yet; current MARIO support is still limited to the Eora 1 style workflows covered here.

Eora26 expects a directory like:

EORA/
├── Eora26_2017_bp/
│   ├── Eora26_2017_bp_T.txt
│   ├── Eora26_2017_bp_FD.txt
│   ├── Eora26_2017_bp_VA.txt
│   ├── Eora26_2017_bp_Q.txt
│   └── Eora26_2017_bp_QY.txt
└── indices/
    ├── labels_T.txt
    ├── labels_FD.txt
    ├── labels_VA.txt
    └── labels_Q.txt

For single-region workflows, path can point to one file or to one directory of files named like IO_ITA_2015_BasicPrice.txt. Use country= and price= when the directory contains multiple candidates.

import mario
Parse one `Eora26` directory

This is the supported multi-region Eora1 26 workflow. Replace the year in path to parse another verified local year. For the local 2016 copy, also pass indeces="/path/to/Eora26_2017_bp".

db = mario.parse_eora(
    path="/path/to/Eora26_2017_bp",
    multi_region=True,
)

db
Parse one local single-region file

This is the non-multi-region Eora workflow. When path points to a directory, country= selects the file to parse.

db = mario.parse_eora(
    path="/path/to/IO_ITA_2017_BasicPrice.txt",
    multi_region=False,
)

db
Select one price variant in the single-region workflow

Use price= when the directory contains more than one variant for the same country and year.

db = mario.parse_eora(
    path="/path/to/IO_All_2017",
    multi_region=False,
    country="ITA",
    price="BasicPrice",
)

db
Aggregate trade rows in the single-region workflow

Use aggregate_trade=True when you want imports and exports collapsed into totals instead of preserving bilateral trade labels.

db = mario.parse_eora(
    path="/path/to/IO_All_2017",
    multi_region=False,
    country="ITA",
    price="BasicPrice",
    aggregate_trade=True,
)

db
Notes and caveats

  • there is no parser here for the full Eora MRIO release;

  • multi-region means Eora26 only;

  • the Eora26 parser applies a few consistency repairs during parsing; those notes are stored in metadata;

  • single-region parsing can infer IOT versus SUT automatically from the local file structure.

db.meta_history

db.sets

db.units

Continue with Inspect and query a database or Calculate matrices.