BEA parser walkthrough#

This notebook shows how to parse the official BEA SUPPLY-USE workbook family with mario.parse_bea(...).

Where the files come from#

Relevant official pages:

Industry Input-Output Accounts Data
Guide to the interactive industry input-output accounts tables
Direct ``SUPPLY-USE.zip` bundle <https://apps.bea.gov/industry/release/zip/SUPPLY-USE.zip>`__

MARIO currently supports only the SUPPLY-USE bundle. It does not yet parse:

MAKE-USE-IMPORTS (BEFORE REDEFINITIONS)
TOTAL AND DOMESTIC REQUIREMENTS

Supported levels and years#

The parser currently supports three aggregation levels:

summary: 1997 to 2024
sector: 1997 to 2024
detail: 2007, 2012, 2017

These ranges were verified directly on the current official workbook bundle.

Accepted path inputs#

parse_bea(...) accepts any of these:

the official SUPPLY-USE.zip
one extracted directory containing the BEA workbooks
one workbook path inside that directory

In all cases you still provide:

year=
level=

Expected extracted structure and matrix mapping#

Typical extracted SUPPLY-USE bundle:

SUPPLY-USE/
├── Supply_Summary.xlsx
├── Use_Summary.xlsx
├── Supply_Sector.xlsx
├── Use_Sector.xlsx
├── Supply_Detail.xlsx
└── Use_SUT_Detail.xlsx

level= selects the workbook pair and year= selects one year inside that pair.

MARIO reads the bundle as a split-native SUT: S from the domestic-industry block of the supply workbook, U and Yc from the use workbook, Va from the use-workbook footer rows, and Vc from the supply-side columns such as imports, margins, and product taxes. The result should therefore be treated as a mixed-valuation SUT: basic-price supply and purchaser-price use.

[1]:

import mario

[2]:

db = mario.parse_bea(
    path="/path/to/SUPPLY-USE",
    year=2024,
    level="summary",
)

db

INFO Parser: BEA SUT payload ready with shapes S=(71, 73), U=(73, 71), Yc=(73, 19), Va=(11, 71), Vc=(11, 73).
INFO Metadata: initialized.

[2]:

name = BEA Supply-Use 2024 Summary
table = SUT
tech_assumption = industry-based
scenarios = ['baseline']
Activity = 71
Commodity = 73
Factor of production = 11
Satellite account = 1
Consumption category = 19
Region = 1

[3]:

db = mario.parse_bea(
    path="/path/to/SUPPLY-USE",
    year=2024,
    level="sector",
)

INFO Parser: BEA SUT payload ready with shapes S=(15, 17), U=(17, 15), Yc=(17, 5), Va=(11, 15), Vc=(11, 17).
INFO Metadata: initialized.

[4]:

db = mario.parse_bea(
    path="/path/to/SUPPLY-USE",
    year=2017,
    level="detail",
)

INFO Parser: BEA SUT payload ready with shapes S=(402, 402), U=(402, 402), Yc=(402, 19), Va=(10, 402), Vc=(10, 402).
INFO Metadata: initialized.

What the parser builds#

MARIO reads the BEA bundle as a split-native SUT:

S from the BEA Supply workbook
U and Yc from the BEA Use workbook
Va from the use-workbook footer rows
Vc from the supply-workbook commodity-side columns for imports, CIF/FOB adjustments, trade and transport margins, and product taxes

This is therefore a mixed-valuation table:

supply is read at basic prices
use is read at purchaser prices

[5]:

base = db["baseline"]
for name in ["S", "U", "Yc", "Va", "Vc"]:
    print(name, base[name].shape)

S (402, 402)
U (402, 402)
Yc (402, 19)
Va (10, 402)
Vc (10, 402)

Caveats#

table="SUT" is the only supported mode right now.
The parser does not yet expose the separate MAKE-USE-IMPORTS (BEFORE REDEFINITIONS) bundle.
The parser does not treat TOTAL AND DOMESTIC REQUIREMENTS as raw database inputs.
detail coverage is much narrower than summary and sector in the current official release.

Download this notebook