1
0
Fork
You've already forked promise-field-analysis
0
Raw data and methodology for two small analyses published in Ch. 8 of my book Integrative Promise
This repository has been archived on 2025年08月04日. You can view files and clone it, but you cannot make any changes to its state, such as pushing and creating new issues, pull requests or comments.
  • Gnuplot 100%
2023年05月21日 19:05:42 +02:00
nlp-over-time.gnuplot All the data, as published. 2023年05月21日 19:05:42 +02:00
nlp-over-time.pdf All the data, as published. 2023年05月21日 19:05:42 +02:00
nlp-over-time.tsv All the data, as published. 2023年05月21日 19:05:42 +02:00
README.md All the data, as published. 2023年05月21日 19:05:42 +02:00

Text Mining in Scientific Fields

This repository contains the raw data that supports two small analyses that I published in chapter 8 of my book Integrative Promise.

Scopus Search

All search results reported here were performed using UCLouvain's Scopus subscription on April 17, 2023.

TITLE-ABS-KEY ( "NLP" OR "natural language processing" OR "text mining" )

This set of keywords is drawn from the methodology at https://royalsocietypublishing.org/doi/10.1098/rspb.2021.2721.

Results: 129,287 documents

Subject Area

Scopus tags each journal with one or more "subject area" tags that describe its content. These are immediately useful for understanding the disciplines that these text mining papers involve. Notably, most of them are published in journals tagged with the subject areas "computer science", "engineering", or "mathematics", disciplines where NLP algorithms are developed directly.

In what follows, I leave these three tags aside. I also do not make any effort to account for the fact that some journals are tagged with multiple subject areas. The

  • Social Sciences (21784)
  • Arts and Humanities (14555)
  • Medicine (12418)
  • Decision Sciences (9770)
  • Business, Management and Accounting (5287)
  • Physics and Astronomy (4157)
  • Biochemistry, Genetics and Molecular Biology (4048)
  • Materials Science (2862)
  • Energy (2454)
  • Health Professions (2323)
  • Environmental Science (2062)
  • Chemical Engineering (1913)
  • Neuroscience (1521)
  • Psychology (1352)
  • Agricultural and Biological Sciences (1231)
  • Multidisciplinary (1188)
  • Chemistry (1078)
  • Economics, Econometrics and Finance (976)
  • Earth and Planetary Sciences (882)
  • Pharmacology, Toxicology and Pharmaceutics (652)
  • Immunology and Microbiology (473)
  • Nursing (425)

The high representation of the social sciences and humanities is quickly attributable to the study of "big data" and social media corpora directly; such journals (like Social Network Analysis and Mining or Scientific Data) are tagged with these tags.

I then group "life" versus "physical" sciences in the following way:

  • Life Sciences (26505)
    • Medicine; Biochemistry, Genetics and Molecular Biology; Health Professions; Environmental Science; Neuroscience; Psychology; Agricultural and Biological Sciences; Pharmacology, Toxicology and Pharmaceutics; Immunology and Microbiology; Nursing
  • Physical Sciences (13346)
    • Physics and Astronomy; Materials Science; Energy; Chemical Engineering; Chemistry; Earth and Planetary Sciences

This figure was used to justify the chapter's claim that the life sciences seem to engage in text mining at a rate about twice as frequent as the physical sciences, and somewhat more frequently than the social sciences.

Graphs over Time

These searches were carried out a few days later, on April 27, 2023.

I then ran the same search above, now limiting it to life sciences and physical sciences (as defined by the sets of subject tags above), respectively. For life science (20,561 documents):

TITLE-ABS-KEY ( "NLP" OR "natural language processing" OR "text mining" ) AND ( LIMIT-TO ( SUBJAREA , "MEDI" ) OR LIMIT-TO ( SUBJAREA , "BIOC" ) OR LIMIT-TO ( SUBJAREA , "HEAL" ) OR LIMIT-TO ( SUBJAREA , "ENVI" ) OR LIMIT-TO ( SUBJAREA , "NEUR" ) OR LIMIT-TO ( SUBJAREA , "PSYC" ) OR LIMIT-TO ( SUBJAREA , "AGRI" ) OR LIMIT-TO ( SUBJAREA , "PHAR" ) OR LIMIT-TO ( SUBJAREA , "IMMU" ) OR LIMIT-TO ( SUBJAREA , "NURS" ) )

For physical sciences (10,209 documents):

TITLE-ABS-KEY ( "NLP" OR "natural language processing" OR "text mining" ) AND ( LIMIT-TO ( SUBJAREA , "PHYS" ) OR LIMIT-TO ( SUBJAREA , "MATE" ) OR LIMIT-TO ( SUBJAREA , "ENER" ) OR LIMIT-TO ( SUBJAREA , "CENG" ) OR LIMIT-TO ( SUBJAREA , "CHEM" ) OR LIMIT-TO ( SUBJAREA , "EART" ) )

Selecting "View All" under "Year" for these searches allowed me to plot these queries over time, used to create the graph figure present here and published in the book. I ignored the data for 2023, as the year was in progress.

The TSV file in this folder contains that data, with years in column 1, life science in column 2, and physical science in column 3.