search

delwen / screenit-tool-comparison

0 stars 0 forks source link

Pull request from screen it report branch #4

Closed elisabascunan closed 4 months ago

elisabascunan commented 4 months ago

This branch is ahead of main by one commit that includes:

1) the updated version of the descriptive-analysis.qmd file, with updated tables and 2 new analyses (comparation table of all the tools plus location analyses of false positives.

2) addition of the two datasets (regset.csv and nc_man_check.csv) to the folder data.

elisabascunan commented 4 months ago

@delwen the changes are ready and committed to the screen it report branch. Ready for pull.

delwen commented 4 months ago

thanks @elisabascunan! The changes look great.

One last comment: in section 5 (description of each tool) just for Sciscore. Since this is the only tool that searches a specific section of the paper, I would add a breakdown of the counts for the locations where an ID was found. So something like the code I wrote below to show the overlap across locations and to drive home the differences for the methods section.

sciscore <- data |> filter(sciscore_hit)
sciscore |> count(id_in_abstract, id_in_methods, id_in_other_location)

id_in_abstract id_in_methods id_in_other_location     n
FALSE          FALSE         TRUE                     2
FALSE          TRUE          FALSE                   28
FALSE          TRUE          TRUE                     3
TRUE           TRUE          FALSE                   21
TRUE           TRUE          TRUE                     3
elisabascunan commented 4 months ago

@delwen very good point! I already made the commit with the table and a short description. Thank you for the observation.