Get data from an image

ocrdata.R

library(dplyr) # for data processing
library(tidyr) # for data processing
library(magick) # for image reading and OCR
library(tesseract) # for OCR

# read image
dataimg <- magick::image_read("https://pbs.twimg.com/media/FBv8P8XXEBgCBvS?format=jpg&name=medium")

# convert image to text
text <- magick::image_ocr(dataimg)

# clean and split text into rows
split_data <- gsub("rs\nCOUNTY VACCINES AL \\|\n\n", "", text) %>%
  gsub("ll. ", "11. ", .) %>% # fix OCR misread
  gsub("\n", " ", .) %>%
  strsplit("\\d+\\. ") %>%
  lapply(trimws)

# split rows into columns and clean
data <- data.frame(x = split_data[[1]][-1]) %>%
  mutate(county = gsub("\\d.*$", "", x),
         number = gsub("^[^0-9]+", "", x)) %>%
  separate(col = number, 
           into = c("number", "percent"), 
           sep = "\\(|%\\)", 
           extra = "drop") %>%
  mutate(number = as.integer(gsub(",", "", number)),
         percent = as.double(percent)) %>%
  select(-x)