Responsible Metrics Dashboard

Code repo including the code to calculate the underlying metrics as well as the Shiny app to present the results.

Forked with many thanks from https://github.com/quest-bih/dashboard

Repo overview

Put all the scripts that are directly involved in preparing data for use in the dashboard in the prep/ folder. Please indicate what scripts/software/data sources are used to generate the inputs for these in a comment in the code itself.

Put the final output data of these scripts in the shiny_app/data/ folder. These should only include the data that the dashboard Shiny app uses directly to generate plots.

The shiny_app/ folder contains the code that is run to generate the Shiny app.

Data preparation pipeline

1. Original data set based on Web of Science and Dimensions search

The following columns came from a search of Web of Science and Dimensions

Column	Type	Details
`doi`	string	Digital Object Identifier
`city`	string	The university medical centre (UMC)
`year_published`	numeric	Year of publication
`early_access`	numeric
`year_searched`	numeric
`email`	string	Corresponding author email
`corr_author`	string	Corresponding author
`authors`	string	Semicolon delimited author list
`affiliations`	string	Author affiliations
`type`	string	Article type
`pmid_wos`	numeric	Pubmed ID (Web of Science)
`wos_categories`	string
`language`	string	Language of article (Web of Science)
`category_for`	json
`mesh_terms`	json
`pmid_dimensions`	numeric	Pubmed ID (Dimensions)
`is_for_match`	boolean
`is_wos_match`	boolean
`is_multidisciplinary`	boolean
`is_book`	boolean
`is_retracted`	boolean
`approach`	categorical	One of three approaches used to id UMC
`specificity_any_TF`	numeric
`specificity_any_TFU`	numeric

2. Open access information

The data set from (1.) is joined by pmid_dimensions with open access information to provide the following additional columns

Column	Type	Details
`color`	categorical	Type of Open Access provided
`issn`	string	International Standard Serial Number
`journal`	string	Name of journal
`publisher`	string	Name of publisher
`date`	date

3. Trial registry number (TRN) reporting

The data set from (2.) is joined by doi with TRN reporting information to provide the following additional columns

Column	Type	Details
`is_human_ct`	boolean	Is this a clinical trial
`abs_trn_1`	string	TRN found in abstract (1)
`abs_trn_2`	string	TRN found in abstract (2)
`abs_trn_3`	string	TRN found in abstract (3)
`abs_trn_4`	string	TRN found in abstract (4)
`abs_trn_5`	string	TRN found in abstract (5)
`abs_trn_6`	string	TRN found in abstract (6)
`abs_registry_1`	string	Registry corresponding to TRN in abstract (1)
`abs_registry_2`	string	Registry corresponding to TRN in abstract (2)
`abs_registry_3`	string	Registry corresponding to TRN in abstract (3)
`abs_registry_4`	string	Registry corresponding to TRN in abstract (4)
`abs_registry_5`	string	Registry corresponding to TRN in abstract (5)
`abs_registry_6`	string	Registry corresponding to TRN in abstract (6)
`si_trn_1`	string	TRN found in secondary information (1)
`si_trn_2`	string	TRN found in secondary information (2)
`si_trn_3`	string	TRN found in secondary information (3)
`si_trn_4`	string	TRN found in secondary information (4)
`si_registry_1`	string	Registry corresponding to TRN in SI (1)
`si_registry_2`	string	Registry corresponding to TRN in SI (2)
`si_registry_3`	string	Registry corresponding to TRN in SI (3)
`si_registry_4`	string	Registry corresponding to TRN in SI (4)

4. Animal research

All unique values from pmid_dimensions column from (3.) are tested for whether they would appear in a search for animal research developed by Hooijmans et al (2010), using https://codeberg.org/bgcarlisle/PubmedIntersectionCheck

The resulting data set is joined by pmid_dimensions to the data set from (3.) to provide the following additional column (see merge-animals-and-sciscore.R for details)

Column	Type	Details
`is_animal`	boolean	Does this paper reflect experimentation on animals

5. Robustness metrics

The data set from (4.) is joined by the pmid_dimensions column with robustness metrics provided by Sciscore to provide the following additional columns (see merge-animals-and-sciscore.R for details)

Column	Type	Details
`sciscore`	numeric	Summary robustness score provided by Sciscore
`iacuc`	boolean	Presence of an IACUC statement
`irb`	boolean	Presence of an IRB statement
`sex`	boolean	Reporting of sex of research subjects
`blinding`	boolean	Reporting of blinding
`randomization`	boolean	Reporting of randomization
`power`	boolean	Presence of a power calculation
`cell_line_auth`
`antibody_detected`
`antibody_with_rrid`
`antibody_rrid_suggested`
`organism_detected`
`organism_with_rrid`
`organism_rrid_suggested`
`tool_detected`
`tool_with_rrid`
`tool_rrid_suggested`
`cell_line_detected`
`cell_line_with_rrid`
`cell_line_rrid_suggested`
`cell_line_contaminated`
`plasmid_detected`
`plasmid_with_rrid`
`plasmid_rrid_suggested`

6. Open code and open data

The data set from (5.) is joined by doi, city and year_published to provide the following additional columns

Column	Type	Description
`success`	boolean
`origin`	categorical
`article`	string	DOI
`is_open_data`	boolean	Whether the article has open data
`open_data_category`	categorical	Location of open data
`is_open_code`	boolean	Whether the article has open code
`open_data_statements`	string	Extracted open data statement
`open_code_statements`	string	Extracted open code statement

7. Prepare UMC names

The UMC names in the city column are now all in lower case. Capitalize them with umc-names-caps.R

bgcarlisle/dashboard