Skip to main content

Scientometrics in R: Bibliometrics, Network Science, and Research Evaluation

6 APIs and Packages

6.1 Learning objectives

After completing this chapter, you will be able to:

  • Query the OpenAlex API using openalexR to retrieve works, authors, institutions, and concepts
  • Apply filters, pagination, and sampling to control the size and scope of API results
  • Deduplicate records by DOI and clean author and affiliation metadata
  • Cache API responses locally so repeated book builds do not re-query the API
  • Use rcrossref and pubmedR as complementary data sources

6.2 Setup 

library(tidyverse)
library(openalexR)
library(glue)

set.seed(20260509)

source(here::here("R", "api_helpers.R"))
source(here::here("R", "utils.R"))
source(here::here("R", "sci_palette.R"))

6.3 Conceptual background

Bibliometric research begins with data. The quality and completeness of a bibliographic corpus determines the validity of every downstream analysis — from citation counts to network maps to topic models.

Historically, Web of Science and Scopus have dominated as data sources for scientometric research. Both require institutional subscriptions and impose restrictions on bulk data use. The launch of OpenAlex in 2022 changed this landscape fundamentally (Priem et al. 2022). OpenAlex is a fully open index of over 250 million scholarly works, built from Crossref, PubMed, institutional repositories, and other open sources. It provides a free REST API with no authentication required, making it the ideal backbone for reproducible bibliometric research.

OpenAlex organises its data around five entity types: works (publications), authors, sources (journals, repositories), institutions, and topics (hierarchical subject tags). Each entity has a unique OpenAlex ID and can be queried via filters on metadata fields such as publication date, citation count, open-access status, and institutional affiliation.

The openalexR package provides a tidy R interface to the OpenAlex API. It handles pagination, rate limiting, and result parsing, returning clean tibbles ready for analysis. For metadata not covered by OpenAlex — such as reference lists and funding information — rcrossref provides access to Crossref’s extensive metadata registry, and pubmedR covers the biomedical literature indexed in PubMed/MEDLINE.

6.4 Worked example

We build a small corpus of recent scientometrics research, demonstrating the full acquisition-to-clean-data pipeline.

6.4.1 Querying works from OpenAlex

The core function is oa_fetch(). We search for works published in the journal Scientometrics between 2020 and 2023, sampling 200 records for a manageable demonstration.

works <- oa_fetch(
  entity = "works",
  primary_location.source.id = "S148561398",
  from_publication_date = "2020-01-01",
  to_publication_date = "2023-12-31",
  options = list(sample = 200, seed = 42)
)

glimpse(works)
#> Rows: 200
#> Columns: 45
#> $ id                            <chr> "https://openalex.org/W4220987815", "htt…
#> $ title                         <chr> "The link between countries’ economic an…
#> $ display_name                  <chr> "The link between countries’ economic an…
#> $ authorships                   <list> [<tbl_df[2 x 7]>], [<tbl_df[2 x 7]>], […
#> $ abstract                      <chr> "Abstract We studied the research perfor…
#> $ doi                           <chr> "https://doi.org/10.1007/s11192-022-0431…
#> $ publication_date              <date> 2022-03-12, 2020-03-07, 2020-02-08, 202…
#> $ publication_year              <int> 2022, 2020, 2020, 2021, 2023, 2020, 2021…
#> $ relevance_score               <dbl> 1.000, 0.999, 0.999, 0.998, 0.997, 0.996…
#> $ fwci                          <dbl> 11.792, 14.757, 4.242, 2.616, 0.474, 1.4…
#> $ cited_by_count                <int> 46, 74, 27, 21, 1, 14, 8, 11, 4, 1, 16, …
#> $ counts_by_year                <list> [<data.frame[5 x 2]>], [<data.frame[7 x…
#> $ ids                           <list> <"https://openalex.org/W4220987815", "h…
#> $ type                          <chr> "article", "article", "article", "articl…
#> $ is_oa                         <lgl> TRUE, TRUE, FALSE, TRUE, FALSE, FALSE, F…
#> $ is_oa_anywhere                <lgl> TRUE, TRUE, FALSE, TRUE, FALSE, FALSE, F…
#> $ oa_status                     <chr> "hybrid", "hybrid", "closed", "hybrid", …
#> $ oa_url                        <chr> "https://link.springer.com/content/pdf/1…
#> $ any_repository_has_fulltext   <lgl> TRUE, TRUE, FALSE, TRUE, FALSE, FALSE, F…
#> $ source_display_name           <chr> "Scientometrics", "Scientometrics", "Sci…
#> $ source_id                     <chr> "https://openalex.org/S148561398", "http…
#> $ issn_l                        <chr> "0138-9130", "0138-9130", "0138-9130", "…
#> $ host_organization             <chr> "https://openalex.org/P4310320108", "htt…
#> $ host_organization_name        <chr> "Springer Nature (Netherlands)", "Spring…
#> $ landing_page_url              <chr> "https://doi.org/10.1007/s11192-022-0431…
#> $ pdf_url                       <chr> "https://link.springer.com/content/pdf/1…
#> $ license                       <chr> "cc-by", "cc-by", NA, "cc-by", NA, NA, N…
#> $ version                       <chr> "publishedVersion", "publishedVersion", …
#> $ referenced_works              <list> <"https://openalex.org/W1581511728", "h…
#> $ referenced_works_count        <int> 66, 63, 45, 74, 76, 63, 74, 72, 51, 64, …
#> $ related_works                 <list> <"https://openalex.org/W4226161467", "h…
#> $ concepts                      <list> [<data.frame[10 x 5]>], [<data.frame[13…
#> $ topics                        <list> [<tbl_df[12 x 5]>], [<tbl_df[12 x 5]>],…
#> $ keywords                      <list> [<data.frame[9 x 3]>], [<data.frame[11 …
#> $ is_paratext                   <lgl> FALSE, FALSE, FALSE, FALSE, FALSE, FALSE…
#> $ is_retracted                  <lgl> FALSE, FALSE, FALSE, FALSE, FALSE, FALSE…
#> $ language                      <chr> "en", "en", "en", "en", "en", "en", "en"…
#> $ sustainable_development_goals <list> [<data.frame[1 x 3]>], NA, [<data.frame…
#> $ awards                        <list> <"https://openalex.org/G1419670400", "G…
#> $ funders                       <list> [<data.frame[2 x 3]>], [<data.frame[3 x…
#> $ apc                           <list> [<data.frame[2 x 5]>], [<data.frame[2 x…
#> $ first_page                    <chr> "2871", "1077", "299", "7917", "4611", "…
#> $ last_page                     <chr> "2896", "1101", "319", "7936", "4650", "…
#> $ volume                        <chr> "127", "123", "123", "126", "128", "124"…
#> $ issue                         <chr> "5", "2", "1", "9", "8", "1", "6", "9", …

The result is a tibble where each row is a work and columns include id, display_name, publication_date, cited_by_count, doi, and nested list-columns for authors and concepts.

6.4.2 Inspecting the results 

works |>
  select(display_name, publication_date, cited_by_count, doi) |>
  arrange(desc(cited_by_count)) |>
  head(10)
#> # A tibble: 10 × 4
#>    display_name                            publication_date cited_by_count doi  
#>    <chr>                                   <date>                    <int> <chr>
#>  1 Publication patterns’ changes due to t… 2021-06-23                  194 http…
#>  2 Link-based approach to study scientifi… 2021-07-10                  192 http…
#>  3 Sample size in bibliometric analysis    2020-07-31                  187 http…
#>  4 Open peer review: promoting transparen… 2020-05-26                  161 http…
#>  5 Identifying interdisciplinary topics a… 2023-07-03                  133 http…
#>  6 Visibility, impact, and applications o… 2023-06-01                  110 http…
#>  7 A bibliometric review of research on i… 2021-04-26                   89 http…
#>  8 Studying the accumulation velocity of … 2020-03-07                   74 http…
#>  9 The impact of research output on econo… 2020-03-28                   71 http…
#> 10 The role of Web of Science publication… 2020-01-02                   69 http…

6.4.3 Querying authors

We can also retrieve author-level data. Let us look up authors associated with scientometrics research by filtering on the relevant topic.

authors <- oa_fetch(
  entity = "authors",
  topics.id = "T10102",
  works_count = ">50",
  options = list(sample = 50, seed = 42)
)

authors |>
  select(display_name, works_count, cited_by_count) |>
  arrange(desc(cited_by_count)) |>
  head(10)
#> # A tibble: 10 × 3
#>    display_name        works_count cited_by_count
#>    <chr>                     <int>          <int>
#>  1 Matthew J. Page             334         210516
#>  2 Daniël Lakens               416          28704
#>  3 Simine Vazire               292          24588
#>  4 Michelle Richardson          93          13568
#>  5 Ernest H. O’Boyle           108          12040
#>  6 Hans‐Dieter Daniel          189           8118
#>  7 Charles Oppenheim           482           7539
#>  8 Jochen Gläser               262           6968
#>  9 Donna K. Ginther            143           6199
#> 10 Mark Oromaner               105           4984

6.4.4 Deduplication by DOI

Real-world corpora often contain duplicates when merging results from multiple queries or sources. The dedupe_by_doi() function from our companion package handles this cleanly.

works_with_dupes <- bind_rows(works, works[1:10, ])
cat(glue("Before dedup: {nrow(works_with_dupes)} rows\n\n"))
#> Before dedup: 210 rows
works_clean <- dedupe_by_doi(works_with_dupes)
cat(glue("After dedup:  {nrow(works_clean)} rows\n"))
#> After dedup:  200 rows

6.4.5 Caching for reproducible builds

The fetch_openalex() wrapper in R/api_helpers.R automatically caches API responses to disk. On subsequent calls with the same arguments, it reads from cache instead of hitting the API.

works_cached <- fetch_openalex(
  entity = "works",
  primary_location.source.id = "S148561398",
  from_publication_date = "2023-01-01",
  to_publication_date = "2023-06-30",
  options = list(sample = 50, seed = 42),
  cache_dir = "_freeze/openalex_cache",
  cache_days = 30
)

nrow(works_cached)
#> [1] 50

6.4.6 Complementary sources: Crossref

For detailed metadata (reference lists, funders, licenses), Crossref is invaluable.

# Requires network access — cached output used in book builds
cr_work <- rcrossref::cr_works(doi = "10.1007/s11192-017-2300-7")
cr_work$data |>
  select(title, container.title, deposited, is.referenced.by.count)

6.4.7 Visualization 

works |>
  mutate(year = year(publication_date)) |>
  count(year) |>
  ggplot(aes(x = year, y = n)) +
  geom_col(fill = palette_sci(1)) +
  labs(x = "Publication year", y = "Number of works") +
  theme_sci()
Bar chart showing the number of sampled publications per year from the journal Scientometrics, covering 2020 to 2023.

Figure 6.1: Annual publication counts for sampled Scientometrics articles (2020–2023).

6.5 Diagnostics and interpretation

When working with API data, always verify:

  • Completeness: Compare your result count against the API’s reported meta$count to confirm you received all expected records. With sampling, the count reflects the sample, not the full set.
  • Coverage dates: Check min(works$publication_date) and max(works$publication_date) to confirm your date filters were applied correctly.
  • DOI presence: Not all works have DOIs. Check mean(!is.na(works$doi)) to understand the DOI coverage rate in your corpus.
  • Nested columns: Author and concept columns are list-columns. Use tidyr::unnest() carefully and inspect for NULL entries before unnesting.
cat(glue("Total works: {nrow(works)}\n"))
#> Total works: 200
cat(glue("DOI coverage: {scales::percent(mean(!is.na(works$doi)))}\n"))
#> DOI coverage: 100%
cat(glue("Date range: {min(works$publication_date)} to {max(works$publication_date)}\n"))
#> Date range: 2020-01-02 to 2023-12-16

6.6 Limitations and responsible use

6.7 Limitations and responsible use

  • Coverage bias: OpenAlex is strongest in English-language, peer-reviewed journal articles. Conference proceedings, books, and non-English literature may be underrepresented. Always report the data source and its known coverage limitations.
  • Metadata quality: Author names and affiliations are parsed algorithmically and may contain errors, especially for names with diacritics or non-Latin scripts.
  • Temporal lag: OpenAlex updates continuously but may lag behind primary databases by days or weeks. Very recent publications may be missing.
  • API rate limits: OpenAlex allows 10 requests per second for unauthenticated users (100/s with a polite pool email). Respect these limits to avoid being throttled.
  • Reproducibility: API results can change as OpenAlex updates its data. Always cache results and document the date of data collection for reproducibility (Hicks et al. 2015).

6.8 Common pitfalls

6.9 Common pitfalls

  • Not caching: Running oa_fetch() directly in code chunks without caching means every quarto render hits the API, producing different results each time and risking rate limits.
  • Ignoring pagination: oa_fetch() handles pagination automatically, but very large queries (>10,000 works) can time out. Use filters to narrow your query or use cursor-based pagination.
  • Confusing sampling with filtering: options = list(sample = 200) returns a random sample from the full result set, not the first 200 records. This is useful for demonstrations but not for exhaustive analyses.
  • Treating OpenAlex IDs as permanent: While generally stable, OpenAlex IDs can change when entities are merged. Use DOIs as the primary identifier for works.
  • Merging sources without deduplication: When combining results from OpenAlex and Crossref, records may overlap. Always deduplicate by DOI before analysis.

6.10 Exercises

  1. Fetch works by institution. Use oa_fetch() to retrieve works affiliated with your own institution. How many works does OpenAlex index for it? What is the DOI coverage rate? (Hint: find your institution’s OpenAlex ID via oa_fetch(entity = "institutions", search = "..."))

  2. Compare Crossref metadata. Pick a DOI from your fetched works and retrieve its metadata from Crossref using rcrossref::cr_works(). Compare the citation count reported by OpenAlex vs. Crossref. Why might they differ?

  3. Build a multi-year corpus. Fetch all works from a journal of your choice for 2015–2023 (without sampling). Plot the annual publication count and the median citation count per year. What trends do you observe?

6.11 Solutions

Solutions are provided in 2.11.

6.12 Further reading

  • Priem et al. (2022) — The OpenAlex paper; describes the data model, coverage, and API design.
  • Aria and Cuccurullo (2017) — The bibliometrix package; an alternative high-level interface for bibliometric data acquisition and analysis.
  • Garfield (1955) — The foundational paper on citation indexing, motivating why these databases exist.

6.13 Session info 

#> R version 4.4.1 (2024-06-14)
#> Platform: x86_64-pc-linux-gnu
#> Running under: Ubuntu 24.04.4 LTS
#> 
#> Matrix products: default
#> BLAS:   /usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so.3 
#> LAPACK: /usr/lib/x86_64-linux-gnu/openblas-pthread/libopenblasp-r0.3.26.so;  LAPACK version 3.12.0
#> 
#> locale:
#>  [1] LC_CTYPE=C.UTF-8       LC_NUMERIC=C           LC_TIME=C.UTF-8       
#>  [4] LC_COLLATE=C.UTF-8     LC_MONETARY=C.UTF-8    LC_MESSAGES=C.UTF-8   
#>  [7] LC_PAPER=C.UTF-8       LC_NAME=C              LC_ADDRESS=C          
#> [10] LC_TELEPHONE=C         LC_MEASUREMENT=C.UTF-8 LC_IDENTIFICATION=C   
#> 
#> time zone: UTC
#> tzcode source: system (glibc)
#> 
#> attached base packages:
#> [1] stats     graphics  grDevices utils     datasets  methods   base     
#> 
#> other attached packages:
#>  [1] rcrossref_1.2.1 gt_1.3.0        tidytext_0.4.3  glue_1.8.1     
#>  [5] openalexR_3.0.1 lubridate_1.9.5 forcats_1.0.1   stringr_1.6.0  
#>  [9] dplyr_1.2.1     purrr_1.2.2     readr_2.2.0     tidyr_1.3.2    
#> [13] tibble_3.3.1    ggplot2_4.0.3   tidyverse_2.0.0
#> 
#> loaded via a namespace (and not attached):
#>  [1] tidyselect_1.2.1   viridisLite_0.4.3  farver_2.1.2       urltools_1.7.3.1  
#>  [5] viridis_0.6.5      S7_0.2.2           fastmap_1.2.0      janeaustenr_1.0.0 
#>  [9] promises_1.5.0     digest_0.6.39      timechange_0.4.0   mime_0.13         
#> [13] lifecycle_1.0.5    tokenizers_0.3.0   brand.yml_0.1.0    magrittr_2.0.5    
#> [17] compiler_4.4.1     rlang_1.2.0        sass_0.4.10        tools_4.4.1       
#> [21] utf8_1.2.6         yaml_2.3.12        knitr_1.51         labeling_0.4.3    
#> [25] stopwords_2.3      htmlwidgets_1.6.4  curl_7.1.0         here_1.0.2        
#> [29] plyr_1.8.9         xml2_1.5.2         RColorBrewer_1.1-3 httpcode_0.3.0    
#> [33] miniUI_0.1.2       withr_3.0.2        triebeard_0.4.1    grid_4.4.1        
#> [37] xtable_1.8-8       scales_1.4.0       dichromat_2.0-0.1  crul_1.6.0        
#> [41] cli_3.6.6          rmarkdown_2.31     generics_0.1.4     otel_0.2.0        
#> [45] rstudioapi_0.18.0  httr_1.4.8         tzdb_0.5.0         cachem_1.1.0      
#> [49] vctrs_0.7.3        Matrix_1.7-0       jsonlite_2.0.0     bookdown_0.46     
#> [53] hms_1.1.4          jquerylib_0.1.4    codetools_0.2-20   DT_0.34.0         
#> [57] stringi_1.8.7      gtable_0.3.6       later_1.4.8        downlit_0.4.5     
#> [61] pillar_1.11.1      htmltools_0.5.9    R6_2.6.1           rprojroot_2.1.1   
#> [65] evaluate_1.0.5     shiny_1.13.0       lattice_0.22-6     SnowballC_0.7.1   
#> [69] memoise_2.0.1      httpuv_1.6.17      bslib_0.11.0       Rcpp_1.1.1-1.1    
#> [73] gridExtra_2.3      xfun_0.57          fs_2.1.0           pkgconfig_2.0.3
This book was built by the bookdown R package.