pub test

pub_test.r

#' ## What's registered in PUB?
#' 
#' ### Fetch the data
#'  
orgs <- rio::import("pevz_org.csv") %>%
  mutate(year = list(2009:2016)) %>%
  tidyr::unnest(year)
u <- "https://pub.uni-bielefeld.de/publication?limit=1000&q="
tt <- purrr::map2(.x = orgs$id, .y= orgs$year, .f = function(x, y) {
  my_url <- URLencode(paste0(u, "department=", x, " AND year=", y, "&fmt=json"))
  jsonlite::fromJSON(my_url)
})
# listen benennen
names(tt) <- orgs$id
# data frame aus listenelementen gewinnen
my_data <- map_df(tt, `[`, c("year", "_id", "type"), .id = "org")
# externe publikationen herausnehmen
ext_ids <- filings %>% 
  select(`_id`, extern) %>% 
  filter(!is.na(extern)) %>%
  .$`_id`
my_data_unibi <- my_data %>% 
  filter(!`_id` %in% ext_ids)
# backup
rio::export(my_data, "counts_bis.csv")
my_counts <- my_data_unibi %>%
  group_by(org) %>%
  count(year)
# filing siehe unten
my_count_all <- filings %>% 
  select(`_id`, extern, year) %>% 
  filter(is.na(extern)) %>%
  group_by(year) %>% 
  summarise(all = n())
my_counts <- my_counts %>% 
  left_join(my_count_all, by = c("year" = "year")) %>%
  mutate(prop = n / all) %>%
  filter(!year == 2009)

ggplot(my_count_all, aes(year, all)) + geom_bar(stat = "identity") +facet_wrap(~type)
### by type
my_counts <- my_data_unibi %>%
  group_by(type) %>%
  count(year)
# filing siehe unten
my_count_all <- filings %>% 
  select(`_id`, extern, year, type) %>% 
  filter(is.na(extern)) %>%
  group_by(year, type) %>% 
  summarise(all = n())
## wie kommen wir zu filings
library(jsonlite)
baseurl <- "https://pub.uni-bielefeld.de/publication?limit=1000&q=year%3E=2010%20AND%20year%3C2017&fmt=json"
pages <- list()
for(i in seq(1,, by = 1000)){
  mydata <- fromJSON(paste0(baseurl, "&start=", i))
  message("Retrieving page ", i)
  pages[[i+1]] <- mydata
}

#' combine all into one
filings <- rbind_pages(pages)
#' ### Analyse
library(tidyverse)
filings %>% 
  group_by(year) %>%
  summarise(n())