USING NAICS AND SIC CODES TO LOCATE FACILITIES BY INDUSTRY
EJAM helps select regulated sites based on industrial classification, using NAICS or SIC code. Finding the right NAICS and finding all the right sites by NAICS is complicated. Doing so requires understanding the NAICS system and the FRS dataset, and the functions in EJAM that help find or use NAICS codes.
NAICS/SIC categories can be explored in a few ways:
- Key EJAM functions for using NAICS/SIC
- NAICS.com website with extensive information about NAICS and SIC
- EPA FRS Facility Industrial Classification Search tool where you can find facilities based on NAICS or SIC.
- EPA APIs exist that can be used for similar queries.
Some key functions include regid_from_naics(),
latlon_from_naics(), frs_from_naics(),
naics_findwebscrape(), and naics_categories().
These functions can help find EPA FRS sites based on naics codes or
titles. They rely on frs_by_naics (a data.table), and
naics_from_any() for querying by code or title of
category.
Files and dataset examples related to NAICS:
topic = "naics"
cbind(data.in.package = sort(grep(topic, EJAM:::datapack()$Item, value = T)))
cbind(files.in.package = sort(basename(testdata(topic, quiet = T))))Important points:
Note that a very large fraction of all FRS sites (as obtained for use in EJAM) lack NAICS code!
Note that EJAM may query FRS sites differently than the FRS search tool or other query tools would.
Note that (NAICS.com) reports many more businesses for a given 6-digit category than the FRS shows, which might be due to FRS only including EPA-regulated sites but also due to data gaps.
Note the difference between
children = TRUEandchildren = FALSEin EJAM functions likelatlon_from_naics()Note that searching on a 6-digit code misses parent categories you may want. The FRS data on NAICS by site is inconsistent in how many digits are reported for the NAICS (explained below).
A given site might be listed in the FRS as being under one or more NAICS codes of various lengths, such as only a parent code (large grouping), only a detailed code (6-digit), or some combination of codes and their subcategories.
And the same title, like “Petroleum Refineries,” may be assigned by
the NAICS system to the category but also a subcategory, as with codes
32411 and 324110. The function naics_from_any() shows what
codes and title exist in the NAICS system.
Also, certain terms appear in the online description of a NAICS but
not in the title of the NAICS – the function
naics_findwebscrape() helps with those cases, e.g., compare
these:
naics_findwebscrape("cement")
naics_from_any("cement")Compare also these:
naics_findwebscrape("refiner")
# reports "324110" (Petroleum Refineries) and other related industries, but not the 5-digit "32411" (also Petroleum
Refineries).
naics_from_any("refiner")
# reports "324110" and "32411" but not other related industries.Using naics_findwebscrape() finds only the 6-digit codes
that match on title or description, so it would find some codes not
found by naics_from_any() which does not query description,
but could lead to missing some facilities in the sense that the 6-digit
code does not cover the sites listed in FRS under only the 5-digit code
for Petroleum Refineries (not the 6-digit).
It is important to note that searching on a 6-digit code misses parent categories that may include sites you expect to find:
frs_from_naics() used as
frs_from_naics("324110", children = F)[,1:5] finds a few
hundred sites, but it fails to find some sites you would find using
frs_from_naics() used as
frs_from_naics("32411", children = F)[,1:5]
The code example below shows that the FRS dataset has some facilities listed under the 5-digit “32411” code only, some with the 6-digit “324110” code only, and some with both codes:
hasboth = intersect(
frs_from_naics("32411", children = F)[,1:5]$REGISTRY_ID,
frs_from_naics("324110", children = F)[,1:5]$REGISTRY_ID
)
hasonly6digit = setdiff(
frs_from_naics("32411", children = F)[,1:5]$REGISTRY_ID,
frs_from_naics("324110", children = F)[,1:5]$REGISTRY_ID
)
hasonly5digit = setdiff(
frs_from_naics("324110", children = F)[,1:5]$REGISTRY_ID,
frs_from_naics("32411", children = F)[,1:5]$REGISTRY_ID
)
length(hasonly5digit) # Most of the FRS sites here
#> [1] 362
length(hasonly6digit)
#> [1] 12
length(hasboth)
#> [1] 12Examples of some NAICS/SIC functions
naics_from_any(naics_categories(3))[order(name),.(name,code)][1:10,]
naics_from_any(naics_categories(3))[order(code),.(code,name)][1:10,]
naics_from_code(211)
naicstable[code==211,]
naics_subcodes_from_code(211)
naics_from_code(211, children = TRUE)
naicstable[n3==211,]
NAICS[211][1:3] # wrong
NAICS[NAICS == 211]
NAICS["211 - Oil and Gas Extraction"]
naics_from_any("plastics and rubber")[,.(name,code)]
naics_from_any(326)
naics_from_any(326, children = T)[,.(code,name)]
naics_from_any("plastics", children=T)[,unique(n3)]
naics_from_any("pig")
naics_from_any("pig ") # space after g
# naics_from_any("copper smelting")
# naics_from_any("copper smelting", website_scrape=TRUE)
# browseURL(naics_from_any("copper smelting", website_url=TRUE) )
a = naics_from_any("plastics")
b = naics_from_any("rubber")
fintersect(a,b)[,.(name,code)] # a AND b
funion(a,b)[,.(name,code)] # a OR b
naics_subcodes_from_code(funion(a,b)[,code])[,.(name,code)] # plus children
naics_from_any(funion(a,b)[,code], children=T)[,.(name,code)] # same
NROW(naics_from_any(325))
#[1] 1
NROW(naics_from_any(325, children = T))
#[1] 54
NROW(naics_from_any("chem"))
#[1] 20
NROW(naics_from_any("chem", children = T))
# [1] 104