Need to extract tables from a pdf using R

General
, ,
4

Here's some code that I believe does what you're looking for. Often when using tabulizer I find I have to manually define the areas of the tables I want to extract. A great way to do this is with the locate_areas() function that will let you interactively drag a box around a table and give you the coordinates to plug into the area argument of extract_tables.

library(tidyverse)
library(tabulizer)

gti_table <- extract_tables(
  "http://visionofhumanity.org/app/uploads/2017/11/Global-Terrorism-Index-2017.pdf",
  output = "data.frame",
  pages = c(106, 106, 107, 107), # include pages twice to extract two tables per page
  area = list(
    c(182, 38, 787, 287),
    c(182, 298, 787, 543),
    c(78, 48, 781, 298),
    c(78, 308, 643, 558)
    ),
  guess = FALSE
  )

gti_table_clean <- reduce(gti_table, bind_rows) %>% # bind elements of list to 1 df
  as_tibble() %>%
  filter(!(X %in% c("GTI RANK", ""))) %>% # remove rownames and empty rows
  rename(
    gti_rank = X,
    county = X.1,
    gti_score = X.2,
    change_score = CHANGE.IN
    ) %>%
  mutate_at(vars(gti_rank, gti_score, change_score), as.numeric) %>% # convert to numeric 
  print()
#> # A tibble: 163 x 4
#>    gti_rank county      gti_score change_score
#>       <dbl> <chr>           <dbl>        <dbl>
#>  1        1 Iraq            10          -0.04 
#>  2        2 Afghanistan      9.44        0.004
#>  3        3 Nigeria          9.01        0.305
#>  4        4 Syria            8.62       -0.033
#>  5        5 Pakistan         8.4         0.214
#>  6        6 Yemen            7.88        0.198
#>  7        7 Somalia          7.65       -0.106
#>  8        8 India            7.53       -0.049
#>  9        9 Turkey           7.52       -0.777
#> 10       10 Libya            7.26        0.027
#> # … with 153 more rows

Created on 2018-10-27 by the reprex package (v0.2.1)

5 Likes