Friendly URL of this spreadsheet:

https://tinyurl.com/rxivist-further-analysis

Rxivist preprint:

https://doi.org/10.1101/515643

Data on Zenodo:

https://doi.org/10.5281/zenodo.2465689

Question: can relation b/w IF and preprint downloads (as shown in Rxivist preprint) be correlated to downloads pre- and post-publication of the respective journal articles?

(and possibly to their OA-availability?)

Comparable figure for ArXiv (one-dimensional, no IF component)

https://twitter.com/catmacOA/status/1041699581233455104

NB This is not the route explored here, because of limitations of usage of absolute download numbers pre- and post-publication.

Acccording to preprint: 37.563 articles, of which 15.797 published

30 most-frequent journals further analyzed (IF, interval to publication)

According to publications_per_journal: 7576 articles in these journals

According to publication_time_journal: 7653 articles in 30 most-frequent journal (interval times listed for these)

Acccording to downloads_by_months: 18823 articles with monthly download stats (for 1st 12 months) = all articles except 2018 (n=18740)

Of articles with monthly stats (2013-2017), 12040 have been published, 6026 in 30 most-frequent journals

Available data per article in this set (n=6026):

Interval to journal publication (days)

Downloads per month for first 12 months

IF of journal (IF for 2017)

OA availability of journal article <- retrieve from UPW met DOI: NO

OA policy of 30 most-frequent journals

Data processing (data_aggregated)

- copy download_publication_status

- VLOOKUP in publication_time_journal for journal name and interval

- interval_month = IFERROR(QUOTIENT(interval,(365/12)),"-") NB This gives the number of full months prior to publication.

- VLOOKUP in downloads_by_months for monthly download stats (via helper columns with 2 values concatenated)

- calculate % prepub and postpub downloads in 1st year

Data processing part 2 (data_aggregated_2)

- copy download_publication_status

- calculate number of articles from pivot data_aggregated

- calculate median_interval, median_downloads_total, median_downloads_year1 and median_prepub directly from data_aggregated, copy into data_aggregated_2

- calculate median_interval_months by dividing median_interval by 365/12

Data processing part 3 (Chart 8 boxplots - data)

- from data_aggregated, calculate 5 number summary (median, Q1, Q3, Q1-1.5*IQR, Q3+1.5*IQR) for % downloads post-publication per journal

- using Google sheets formulas MEDIAN, QUARTILE

- Template for making box plots taken from:

https://creativemaths.net/blog/teaching-resources-item/google-sheets-compare-box-plots/

List of charts in this spreadsheet:

IF vs. downloads

IF vs. downloads year 1

IF vs. interval until publication

interval until publication vs. % downloads post-publication

IF vs. % downloads post-publication

downloads year 1 vs. % downloads post-publication

scatterplots per journal (downloads year 1 vs. % downloads post-publication)

boxplots (% downloads post-publication)

Excel sheet with underlying data:

https://www.dropbox.com/s/69lxaab17u2efld/Rxivist%20analysis.xlsx?dl=0

created 190123

last modified 190127

original data shared by the authors of the Rxivist preprint under CC-BY-NC