FIGARO

Import Eurostat’s inter-country supply-use or input-output tables.

Download notebook

The parser uses Eurostat data and a local cache. Select table type and, for IOTs, the product or industry system.

Import#

import mario

db = mario.parse_figaro(table="SUT", year=2022, path="figaro_cache")

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

Inputs and options#

Source links

FIGARO IOT values are in million EUR at current basic prices. Products follow CPA 2.1 and activities NACE Rev.2. Browser monetary values are converted to current USD using same-year exchange rates. Derived calculations must credit Eurostat; Eurostat is not responsible for these calculations.

Dataflow groups

The parser selects the dataflow suffix from year:

2010-2013: s1, u1, ip1, ii1
2014-2017: s2, u2, ip2, ii2
2018-2021: s3, u3, ip3, ii3
2022 onwards: s4, u4, ip4, ii4

For example, a 2022 SUT uses naio_10_fcp_s4 and naio_10_fcp_u4.

Supply flows use a single geo dimension: MARIO places domestic production on the same region on both axes of S. Use and IOT flows retain their origin and destination dimensions.

Key arguments

table: choose "SUT" or "IOT".
year: reference year, required; this is separate from the dataset edition.
path: optional directory of normalized API-cache CSVs and JSON sidecars. Matching caches are reused; missing requests are downloaded into that directory.
countries: optional FIGARO country-code selection; keep it consistent with the cached request. A subset is not a complete global MRIO.
iot_mode: "product" → pxp; "industry" → ixi; "auto" defaults to product. This is an IOT system, separate from table format.
unit: Eurostat monetary unit, currently "MIO_EUR".

Local caches and API requests

MARIO checks path= for a matching cache before calling the API. The cache key includes the dataflow, year, monetary unit, country selection and factor-row selection; keep the original cache filenames. path is a cache directory, not a single CSV path.

Without a matching cache, MARIO chunks requests by c_orig to respect Eurostat extraction limits. Full tables are large, so country subsets can help during development; they do not replace the full global parse used in the browser.

The DOM/W2 rows supply the factor matrix V. For GDP, sum D1 + D29X39 + B2A3G and add net product taxes D21X31 on intermediate and final uses. The IOT parser retains intermediate taxes in V; the browser independently reads final-use taxes from the same source cache. Residence-purchase adjustment rows are excluded from this income calculation.

import mario

Parse a FIGARO SUT

Point path to the directory containing matching supply and use API caches. If either cache is absent, MARIO downloads that request.

db_sut = mario.parse_figaro(
    table="SUT",
    year=2022,
    path="/path/to/figaro/sut_directory"
)

db_sut

db_sut

Parse a FIGARO IOT

The local series uses iot_mode="product" (pxp). Change year and select its cache directory to read another annual table. industry (ixi) is supported by MARIO but is not present in the supplied local IOT inventory.

db_iot = mario.parse_figaro(
    table="IOT",
    year=2022,
    iot_mode="product",
    path="/path/to/figaro/iot_directory",
    calc_all=False,
)

db_iot

First inspection

Once parsed, the result is a standard MARIO database.

db_sut.get_index("Region")[:5]

db_sut.get_index("Activity")[:5]

db_sut.get_index("Commodity")[:5]

Practical notes

ip1–ip4 and ii1–ii4 partition reference years; their suffixes are not database versions.
Preserve cache filenames and the original country selection. countries= changes the request key.
The public browser groups identical native sets into one Excel inventory; equal coverage does not mean equal annual values.
No GHG indicator is inferred from monetary factors or MARIO placeholder extension rows. Environmental accounts require separate source files.

Continue with Inspect and query a database or Calculate matrices.