econiodatajp provides ready-to-use Japanese input-output (IO) tables,
built from data published by the Ministry of Internal Affairs and
Communications (e-Stat) and other sources.
Each table is returned as an
econio input-output table
object, so it works directly with econio’s analysis functions
(io_reclass(), io_leontief_inverse(), and so on).
Tables are archived and rebuilt on demand via
tarchives/
targets, so io_table_get() (or
io_table_target(), for use inside your own _targets.R pipeline)
downloads and processes each source file only once per version.
You can install the development version of econiodatajp from GitHub with:
# install.packages("pak")
# pak::pak("UchidaMizuki/econiodatajp")
install.packages("econiodatajp", repos = c('https://uchidamizuki.r-universe.dev', 'https://cloud.r-project.org'))
io_table_get() covers three shapes of table, selected via
region_type and region_class. sector_class has no default and must
always be specified; language defaults to "ja" (the original
Japanese sector names) and only accepts "en" for the national table.
The national IO table (region_type = "regional" and
region_class = "nation" are both the default):
library(econiodatajp)
io_table_get(year = 2020, sector_class = "medium")
#> trying URL 'https://www.e-stat.go.jp/stat-search/file-download?statInfId=000040187024&fileKind=0'
#> downloaded 734 KB
#>
#> trying URL 'https://www.e-stat.go.jp/stat-search/file-download?statInfId=000040186856&fileKind=0'
#> downloaded 68 KB
#>
#> trying URL 'https://www.e-stat.go.jp/en/stat-search/file-download?statInfId=000040186856&fileKind=0'
#> downloaded 59 KB
#>
#> # Input-output table: regional
#> # Dimensions: input [115], output [119]
#> # Input: 115 sectors
#> # Output: 119 sectors
#> # Import type: competitive
#> input$sector output$sector
#> <sector> <sector>
#> 1 <industry> 011_耕種農業 <industry> 011_耕種農業
#> 2 <industry> 011_耕種農業 <industry> 012_畜産
#> 3 <industry> 011_耕種農業 <industry> 013_農業サービス
#> 4 <industry> 011_耕種農業 <industry> 015_林業
#> 5 <industry> 011_耕種農業 <industry> 017_漁業
#> 6 <industry> 011_耕種農業 <industry> 061_石炭・原油・天然ガス
#> 7 <industry> 011_耕種農業 <industry> 062_その他の鉱業
#> 8 <industry> 011_耕種農業 <industry> 111_食料品
#> 9 <industry> 011_耕種農業 <industry> 112_飲料
#> 10 <industry> 011_耕種農業 <industry> 113_飼料・有機質肥料（別掲を除く。）
#> # ℹ 13,675 more rows
#> # ℹ 1 more variable: . <dbl>
Pass language = "en" for English sector names:
io_table_get(year = 2020, sector_class = "medium", language = "en")
#> # Input-output table: regional
#> # Dimensions: input [115], output [119]
#> # Input: 115 sectors
#> # Output: 119 sectors
#> # Import type: competitive
#> input$sector
#> <sector>
#> 1 <industry> 011_Crop cultivation
#> 2 <industry> 011_Crop cultivation
#> 3 <industry> 011_Crop cultivation
#> 4 <industry> 011_Crop cultivation
#> 5 <industry> 011_Crop cultivation
#> 6 <industry> 011_Crop cultivation
#> 7 <industry> 011_Crop cultivation
#> 8 <industry> 011_Crop cultivation
#> 9 <industry> 011_Crop cultivation
#> 10 <industry> 011_Crop cultivation
#> # ℹ 13,675 more rows
#> # ℹ 2 more variables: output <tibble[,1]>, . <dbl>
One prefecture (region_class = "pref" with region_type = "regional"
requires region, the "NN_name" code_name fragment used in the
archive); sector_class is that prefecture’s own sector count, not a
tier shared across prefectures – use io_table_available() to look
both up:
io_table_get(region_class = "pref", region = "01_hokkaido", year = 2015, sector_class = "105")
#> trying URL 'https://www.hkd.mlit.go.jp/ky/ki/keikaku/splaat000001yqxt-att/splaat000001yr7c.xlsx'
#> Content type 'application/vnd.openxmlformats-officedocument.spreadsheetml.sheet' length 1637807 bytes (1.6 MB)
#> ==================================================
#> downloaded 1.6 MB
#>
#> # Input-output table: regional
#> # Dimensions: input [112], output [117]
#> # Input: 112 sectors
#> # Output: 117 sectors
#> # Import type: competitive
#> input$sector output$sector .
#> <sector> <sector> <dbl>
#> 1 <industry> 01_食用耕種農業 <industry> 01_食用耕種農業 20647000000
#> 2 <industry> 01_食用耕種農業 <industry> 02_非食用耕種農業 1217000000
#> 3 <industry> 01_食用耕種農業 <industry> 03_畜産 6907000000
#> 4 <industry> 01_食用耕種農業 <industry> 04_農業サービス 732000000
#> 5 <industry> 01_食用耕種農業 <industry> 05_林業 99000000
#> 6 <industry> 01_食用耕種農業 <industry> 06_漁業 0
#> 7 <industry> 01_食用耕種農業 <industry> 07_石炭・原油・天然ガス 0
#> 8 <industry> 01_食用耕種農業 <industry> 08_その他の鉱業 0
#> 9 <industry> 01_食用耕種農業 <industry> 09_畜産食料品 195000000
#> 10 <industry> 01_食用耕種農業 <industry> 10_水産食料品 563000000
#> # ℹ 13,094 more rows
Every prefecture at once, as a single table with a region dimension
(region isn’t used for a "multiregional" table – it always covers
every region in the breakdown at once):
io_table_get(
region_type = "multiregional",
region_class = "pref",
year = 2011,
sector_class = "31"
)
#> trying URL 'https://www.rieti.go.jp/jp/database/r-io2011/data/i-preio2011.xlsx'
#> Content type 'application/vnd.openxmlformats-officedocument.spreadsheetml.sheet' length 14015943 bytes (13.4 MB)
#> ==================================================
#> downloaded 13.4 MB
#>
#> # Input-output table: multi-regional
#> # Dimensions: input [1,739], output [1,833]
#> # Input: 47 regions, 31 industries
#> # Output: 47 regions, 31 industries
#> # Import type: competitive
#> input$region $sector output$region .
#> <glue> <sector> <glue> <dbl>
#> 1 01_北海道 <industry> 0100_農林水産業 01_北海道 242716000000
#> 2 01_北海道 <industry> 0100_農林水産業 01_北海道 14000000
#> 3 01_北海道 <industry> 0100_農林水産業 01_北海道 631181000000
#> 4 01_北海道 <industry> 0100_農林水産業 01_北海道 165000000
#> 5 01_北海道 <industry> 0100_農林水産業 01_北海道 804000000
#> 6 01_北海道 <industry> 0100_農林水産業 01_北海道 279000000
#> 7 01_北海道 <industry> 0100_農林水産業 01_北海道 11000000
#> 8 01_北海道 <industry> 0100_農林水産業 01_北海道 0
#> 9 01_北海道 <industry> 0100_農林水産業 01_北海道 0
#> 10 01_北海道 <industry> 0100_農林水産業 01_北海道 0
#> # ℹ 3,187,577 more rows
#> # ℹ 1 more variable: output$sector <sector>
Or the official 9-region block breakdown instead
(region_class = "block"; published for 1970, 1975, 1980, 1985, 1990,
1995, and 2005 – use io_table_available() to list what’s on hand):
io_table_get(
region_type = "multiregional",
region_class = "block",
year = 2005,
sector_class = "12"
)
#> trying URL 'https://www.e-stat.go.jp/stat-search/file-download?statInfId=000020467390&fileKind=0'
#> downloaded 3.6 MB
#>
#> # Input-output table: multi-regional
#> # Dimensions: input [180], output [171]
#> # Input: 9 regions, 12 industries
#> # Output: 9 regions, 12 industries
#> # Import type: competitive
#> input$region $sector output$region .
#> <glue> <sector> <glue> <dbl>
#> 1 1_北海道 <industry> 01_農林水産業 1_北海道 274730000000
#> 2 1_北海道 <industry> 01_農林水産業 1_北海道 180000000
#> 3 1_北海道 <industry> 01_農林水産業 1_北海道 659752000000
#> 4 1_北海道 <industry> 01_農林水産業 1_北海道 0
#> 5 1_北海道 <industry> 01_農林水産業 1_北海道 1000000
#> 6 1_北海道 <industry> 01_農林水産業 1_北海道 41690000000
#> 7 1_北海道 <industry> 01_農林水産業 1_北海道 2807000000
#> 8 1_北海道 <industry> 01_農林水産業 1_北海道 0
#> 9 1_北海道 <industry> 01_農林水産業 1_北海道 252000000
#> 10 1_北海道 <industry> 01_農林水産業 1_北海道 4000000
#> # ℹ 30,770 more rows
#> # ℹ 1 more variable: output$sector <sector>
sector_class controls the sector classification granularity
("basic", "small", "medium", "large", or "template" for
region_class = "nation", each a direct translation of a real official
Japanese classification tier name). No official source names the sector
granularities of a region_class = "pref" or "block" table, only
counts them, so sector_class there is that count as a string instead:
each prefecture’s own count for its region_type = "regional" table
(e.g. "105" for Hokkaido); "26"/"31" (2005/2011) for the
region_class = "pref" multiregional table; and "12"/"29"/"53" in
2005, or a single year-specific count otherwise (e.g. "43" in 1970),
for the region_class = "block" multiregional table – use
io_table_available() to look any of these up. import_type and
language are only meaningful for region_class = "nation".
To declare one of these tables as a target in your own {targets}
pipeline instead of fetching it eagerly, use io_table_target() with
the same arguments:
# _targets.R
library(targets)
library(econiodatajp)
list(
io_table_target(iotable_nation_2020, year = 2020, sector_class = "medium")
)