Converts an sm_corpus to the data frame format used by
bibliometrix::biblioAnalysis() and the other functions of bibliometrix:
one row per work and one column per Web of Science field tag. This
enables using bibliometrix's analysis functions on corpora built with
scimapR.
Usage
sm_to_bibliometrix(
corpus,
keep_extra = TRUE,
db = c("auto", "ISI", "scimapR"),
...
)Arguments
- corpus
An
sm_corpusobject.- keep_extra
Logical. If
TRUE(default), the columns without a field tag follow the tag columns, as in earlier versions: the otherworkscolumns (such aswork_idandsource_id),country_codeandSO_ISSN.FALSEreturns the tag columns only.- db
Value of the
DBcolumn."auto"(default) writes"ISI"when every work comes from Web of Science and"scimapR"otherwise;"ISI"and"scimapR"write that value. See Details.- ...
Additional arguments (currently unused).
Value
A data frame with class c("bibliometrixDB", "data.frame"), one
row per work in the order of corpus$works, with SR as row names.
Every tag column is present, NA where the corpus holds no data.
Details
Field tags
AU: authors asSURNAME INITIALS,AF: full names asSURNAME, GIVEN NAMES; one entry per author of the work, inpositionorder and separated by;, fromauthors$display_namethroughauthorships. An author with several address rows appears once; authors without a name are left out.TI,AB:works$title,works$abstract;TI_rawandAB_rawhold them in their stored case, asbibliometrix::convert2df()does.SO,J9,JI:sources$display_name, and the optional columnssources$abbreviationandsources$iso_abbreviation.SN,EI:sources$issn_l(else the first value ofsources$issn) and the next distinct value ofsources$issn.C1: one segment per distinct address of the work,[AUTHOR A; AUTHOR B] ADDRESS, with the authors (AFform) inpositionorder, joined by;; fromauthorships$raw_affiliation.RP: one segment per corresponding author (authorships$is_corresponding),SURNAME, INITIALS (CORRESPONDING AUTHOR), ADDRESSwith the author's first address, joined by;.EM: the e-mail addresses of the optional columnauthorships$email.CR:references$cited_rawinref_indexorder, separated by;;NR: the number of items inCR.DE: author keywords, see below;DE_rawholds them in their stored case.ID: the concepts of theindexer_keywordfamily ofsm_concept_vocabularies()(such as Keywords Plus or MeSH).WC: the vocabularywos-category(Web of Science categories);SC: the vocabularywos-subject-category(research areas). Terms are joined by;.DT,LA,PY,TC,DI,PM:works$type,language,year,cited_by_count,doiandpmid.UT: the Web of Science accession number,provenance$source_id_externalof the work's first provenance row withsource == "wos", else awork_idthat starts withWOS:.DB: see below.SR: the short referenceFIRST AUTHOR, YEAR, J9(withSOwhereJ9is missing), made unique by appending-a,-b, ... to repeats;SR_FULLkeeps it without the suffixes.
With keep_extra = TRUE, country_code holds the country codes of
authorships$country_code, one per authorship row (author and
address), and SO_ISSN holds sources$issn_l. The export has no
AU_CO column, as data from bibliometrix::convert2df() has none:
bibliometrix::metaTagExtraction(M, "AU_CO") adds the country names
that bibliometrix reads from C1 and uses for its country analyses.
Conventions
Like bibliometrix::convert2df() for Web of Science files, the export
upper-cases AU, AF, TI, SO, J9, JI, AB, C1, RP, EM,
CR (including the DOIs in it), DE, ID, WC, SC, DT and LA.
DI, UT, SN, EI and PM are written as stored, so DI keeps the
lower-case DOIs of the scimapR readers. Upper-casing uses toupper(),
whose result for non-ASCII letters depends on the locale.
Names: a display_name with a comma is read as Surname, Given names.
Otherwise an entry of authors$display_name_alternatives in that form
that agrees with the name gives the surname, so a surname of several
words such as "Garcia Lopez" stays whole; sm_read_wos() keeps the
names of the file there. Else initials at the end of the name (single
letters, or capitals with periods, as in "Chen W." from
sm_read_scopus()) follow the surname, and otherwise the last word is
the surname, together with the lower-case particles before it ("Anna
van der Berg" gives VAN DER BERG A). Particles never become initials.
Without such a spelling, a surname of several words without particles
("Maria Garcia Lopez") keeps only its last word, and a name such as
"SMITH JA" is read as given names followed by the surname. The given
names of AF are the longest spelling at hand, the initials of AU
the shortest. Corpora read by sm_read_wos() store the AU form of a
name (such as "J Smith"), so their AF holds initials.
CR items are written in the Web of Science form SURNAME INITIALS, YEAR, SOURCE, ...: an item Surname, IN, YEAR, ... (with a comma after
the surname, as some exports write it) becomes SURNAME IN, YEAR, ....
Other items, such as titles or OpenAlex IDs, are only upper-cased.
Because ; separates the items of AU, AF, C1, RP, CR, DE,
ID, WC and SC, a ; inside an item becomes ,.
DE holds the author keywords of each work: the concepts of the
author_keyword family of sm_concept_vocabularies() (author-keywords,
manual-tags). A work without author keywords gets its OpenAlex
concepts (openalex) instead, so corpora built from OpenAlex keep their
DE; other vocabularies (Keywords Plus, MeSH, categories, arXiv and
bioRxiv subjects) never enter DE. A work with neither has DE = NA.
DB: with db = "auto", a corpus whose works all come from Web of
Science (each has a provenance row with source == "wos", or a
work_id that starts with WOS:) gets "ISI", the value
bibliometrix::convert2df() writes for Web of Science files, so that
bibliometrix::histNetwork() and bibliometrix::rpys() accept it.
Other corpora get "scimapR", because bibliometrix reads "ISI" data
as Web of Science records; db = "ISI" declares a corpus whose CR
items are in Web of Science form anyway.
RP follows authorships$is_corresponding: it names the authors
flagged TRUE and is NA for a work without one. sm_read_wos() takes
the corresponding authors from the RP field of the file,
sm_read_openalex_json() and sm_fetch_openalex() from the OpenAlex
records, sm_read_dimensions() from the Corresponding Authors column,
sm_fetch_biorxiv() from the bioRxiv API and as_sm_corpus() from
RP. The other readers and fetchers record the corresponding author as
unknown (NA), so their corpora export RP = NA. Corpora built from
tables keep the flags of their tables.
as_sm_corpus() reads the result back: the round trip keeps the works,
their DOIs and years, the authors of each work, their addresses and the
corresponding authors.
References
Aria, M. & Cuccurullo, C. (2017). bibliometrix: An R-tool for comprehensive science mapping analysis. Journal of Informetrics, 11(4), 959-975. doi:10.1016/j.joi.2017.08.007
Examples
corpus <- sm_example_corpus(n_works = 10, n_authors = 5,
with_embeddings = FALSE)
M <- sm_to_bibliometrix(corpus, keep_extra = FALSE)
class(M)
#> [1] "bibliometrixDB" "data.frame"
M[1:3, c("AU", "PY", "SO", "DB")]
#> AU PY
#> ZHANG M, 2023, PLOS ONE ZHANG M;PETROVA D;SMITH C;JOHNSON H;FISCHER R 2023
#> JOHNSON H, 2019, PNAS JOHNSON H;SMITH C;ZHANG M 2019
#> FISCHER R, 2018, PLOS ONE FISCHER R;SMITH C;ZHANG M;JOHNSON H 2018
#> SO DB
#> ZHANG M, 2023, PLOS ONE PLOS ONE scimapR
#> JOHNSON H, 2019, PNAS PNAS scimapR
#> FISCHER R, 2018, PLOS ONE PLOS ONE scimapR
# a corpus from Web of Science records
fix <- readRDS(sm_example_files("example_realuse_corpus.rds"))
M <- sm_to_bibliometrix(fix)
unique(M$DB)
#> [1] "ISI"
M$C1[1]
#> [1] "[VAHRENKAMP, SVENJA] ZENT INST VETERINARWESENS NORDSTADT, DEPT OCCUPAT MED, D-99003 NORDSTADT, GERMANY.; [HOEKSTRA, PRIYA] LAKESIDE UNIV, DEPT EPIDEMIOL, LAKESIDE, MA 02999 USA.; [PIEPER, CLARA; NOLTING, MATTHIAS; LUTTERBECK, MARIE] ZENT INST VETERINARWESENS NORDSTADT, DEPT HYG, D-99003 NORDSTADT, GERMANY.; [ENGELKE, CLARA] FED PUBLIC SERVICES INST MICROBIOL, DEPT VIROL, D-99002 SOUTHCITY, GERMANY.; [ENGELKE, CLARA] UNIV NORTHTOWN, DEPT MICROBIOL, NORTHTOWN, GERMANY.; [PETROVIC, HANNAH] SOUTHCITY MED SCH, DEPT INFECT DIS, D-99012 SOUTHCITY, GERMANY."
# \donttest{
if (requireNamespace("bibliometrix", quietly = TRUE)) {
results <- bibliometrix::biblioAnalysis(M, sep = ";")
}
# }