compiles publications & citation list for OPIK output for a given year range; requires some metadata in Drive

compile_citations_OPIK.R

library(scholar)
library(googlesheets4)
library(tidyverse)
gs4_auth(email = "[email protected]")
authors <- read_sheet(ss = "https://docs.google.com/spreadsheets/d/18RWKaMvkUOGp_URLwRu-3jTLUz2iI4y3oN_hu8MLoRg/edit#gid=0")
authors
pubsi <- list()
for (i in 1:nrow(authors)){
  pubsi[[i]] <- get_publications(id = authors$scholar_id[i],
                                 sortby = "year", 
                                 flush = TRUE) %>% 
    dplyr::filter(between(year, authors$desde[i], authors$hasta[i])) %>% 
    mutate(author_id = authors$scholar_id[i])
  Sys.sleep(1)
}
pubsi[[1]] %>% View()
db <-
  pubsi %>% 
  bind_rows() %>% 
  mutate(keep = !duplicated(pubid),
         keep = if_else(journal %in% c("","Figshare"), FALSE, keep),
         keep = if_else(grepl(journal, pattern = "MPIDR"), FALSE, keep),
         keep = if_else(grepl(journal, pattern = "Conference"), FALSE, keep),
         keep = if_else(grepl(journal, pattern = "MPIDR"), FALSE, keep),
         keep = if_else(grepl(journal, pattern = "PAA"), FALSE, keep),
         keep = if_else(grepl(journal, pattern = "demogr.mpg.de"), FALSE, keep),
         keep = if_else(grepl(tolower(journal), pattern = "rxiv"), FALSE, keep),
         keep = if_else(grepl(journal, pattern = "www.mortality.org"), FALSE, keep),
         keep = if_else(grepl(journal, pattern = "https://rdrr."), FALSE, keep),
         keep = if_else(grepl(journal, pattern = "ISA Forum"), FALSE, keep),
         keep = if_else(grepl(journal, pattern = "project-WP"), FALSE, keep),
         keep = if_else(grepl(journal, pattern = "OPIK Working"), FALSE, keep),
         keep = if_else(grepl(journal, pattern = "OPIK-Working"), FALSE, keep),
         keep = if_else(grepl(journal, pattern = "Congreso"), FALSE, keep),
         keep = if_else(grepl(tolower(journal), pattern = "documento de trabajo"), FALSE, keep),
         keep = if_else(grepl(tolower(journal), pattern = "the conversation"), FALSE, keep),
         keep = if_else(grepl(journal, pattern = "Population Association of"), FALSE, keep)) %>% 
  dplyr::filter(keep) %>% 
  arrange(year) %>% 
  mutate(journal = case_when(grepl(tolower(journal), 
                                   pattern = "international journal of integrated care")~
                               "International Journal of Integrated Care",
                             grepl(tolower(journal),
                                   pattern = "plos one") ~ "PLOS ONE",
                             (journal %>% 
                                stringi::stri_trans_general(id = "Latin-ASCII") %>% 
                                tolower()) %in%
                               c("rev esp salud publica", 
                                 "revista espanola de salud publica") ~ "Revista Española de Salud Pública",
                             TRUE ~ journal),
         journal = str_to_title(journal))  
db %>% 
  select(-author_id,
         -keep) %>% 
  write_sheet(ss = "https://docs.google.com/spreadsheets/d/1nR_v_DH-tV-xvcT246UqnC3TJAZBjLDhwpaI8m9Q6B4/edit#gid=0",
              sheet = "resultados")

# from here, many duplicated discovered, which need to be merged in Google Scholar.


db$journal %>% unique()

# start on journal impact statistics. 

# SCIMAGO
# devtools::install_github("ikashnitsky/sjrdata")
# library(sjrdata)
# sjr_journals %>%
#   dplyr::filter(title %in% (db$journal %>% unique()) )