R Notebook

2020 network

Simple scrape of URL and re-ordering of the rows by size

#install.packages('janitor')
library(janitor)

library(rvest)
library(tidyverse)
library(DT)

#scrape the table from url into d variable
url='https://www.ohdsi.org/web/wiki/doku.php?id=resources:2020_data_network'

content <- read_html(url)
tables <- content %>% html_table(fill = TRUE)


d <- tables[[1]]


#fix columns

cols=d  %>% slice(1) %>% unlist() %>% unname()



#reduce cols (only  8 have actual data)
d %<>%  select(1:8) 



#add names
names(d)=cols[1:8]
#names(d)


d %<>% clean_names()
#drop row 1 with col names
d %<>% slice(2:n())

#make the number nicer (replace comma with nothing and convert to number)
#still have some problematic entries
d %<>% mutate(pt_cnt=as.numeric(str_replace_all(number_of_patients,',',''))) %>% arrange(desc(pt_cnt))

NAs introduced by coercion

#drop first column, no value provided
d$number = NULL

List

d %>% datatable()

d$year=2020
d20=d

2019

url='https://www.ohdsi.org/web/wiki/doku.php?id=resources:2019_data_network'
content <- read_html(url)
tables <- content %>% html_table(fill = TRUE)


d <- tables[[1]]


#fix columns

#cols=d  %>% slice(1) %>% unlist() %>% unname()



#reduce cols (only  8 have actual data)
#d %<>%  select(1:8) 



#add names
#names(d)


d %<>% clean_names()
#drop row 1 with col names
#d %<>% slice(2:n())

#make the number nicer (replace comma with nothing and convert to number)
#still have some problematic entries
d %<>% mutate(pt_cnt=as.numeric(str_replace_all(number_of_patients,',',''))) %>% arrange(desc(pt_cnt))

NAs introduced by coercion

#drop first column, no value provided
d$number = NULL

List

d %>% datatable()

d$year=2019
d19=d
alld=bind_rows(d20,d19) %>% arrange(desc(year),desc(pt_cnt))
alld %>% write_csv('ohdsi-data-network-combined.csv')

Combined years

alld %>% datatable(rownames = FALSE)

Analysis by dataset type

alld %>% group_by(year,data_type) %>% summarize(n=n()) %>% arrange(year,desc(n)) %>% datatable(rownames = FALSE)

`summarise()` has grouped output by 'year'. You can override using the `.groups` argument.