How to create a CV with Rmarkdown, Awesome CV and Google Sheets

A reproducible workflow for building and updating a polished CV with R Markdown.

Author

Darwin Cubi

Published

March 2, 2022

CV in R

When entering the working world or looking for new job opportunities we have all been in the task of creating or updating our resume and almost always the main tool is Word, Canva or Adobe Illustrator (for the more advanced). The last two options, although they offer a lot of customization in terms of design, do not allow much flexibility when updating information or easily adapting to a workflow. On the other hand, Word has templates with somewhat bland designs and the constant struggle with alignments, indents and styles when entering new information.

A couple of years ago I was confronted with this problem and I was lucky to find the solution in the vitae package developed by Mitchell O’Hara-Wild. This package allows you to create a cv with 6 types of formatting:

Awesome-CV format
Hydman
ModernCV
Twenty seconds
latexcv
markdown-cv

Here you can see the examples of the different formats together with their source code. Personally I like the Awesome-CV template and in fact this whole tutorial is developed focused on this style of cv.

First you must create a project in Rstudio because we will take advantage of git version control, in this link you can find the step by step of how to create a project with git version control.

The first step is to install the vitae package

install.packages("vitae")

and then create an awesome cv template with markdown

RStudio menu for creating a new R Markdown document R Markdown template dialog with the Awesome CV format selected This will create the following template

RStudio editor showing Awesome CV YAML parameters and example entries To compile you must click on ‘Knit’ and you will have the following pdf

Rendered Awesome CV example for Marie Curie In the YAML parameters you can specify data such as name, position, address, email, etc. Here you can find more details about each available parameter.

The data for the cv is entered in a tribble and transformed to awesome format with the detailed_entries() function that has the parameters:

data: tribble, tibble or dataframe with the data
what: name of the activity or title
when: date of the activity or title
with: company or institution
where: place of the activity or title
why : for additional data (this can be a list of values)

There is also function brief_entries(data, what, when, with, .protect = TRUE) which is similar but with less arguments. For more information about these functions you can type in the console ?detailed_entries or ?brief_entries.

The following code

tribble(
  ~ Degree, ~ Year, ~ Institution, ~ Where,
  "Informal studies", "1889-91", "Flying University", "Warsaw, Poland",
  "Master of Physics", "1893", "Sorbonne Université", "Paris, France",
  "Master of Mathematics", "1894", "Sorbonne Université", "Paris, France"
) %>% 
  detailed_entries(Degree, Year, Institution, Where)

compiling to pdf will create

Rendered education entries generated from structured R data At this point you will have created a functional cv and will not need to worry about layout, indentation or text alignment. But it is still cumbersome to have the data and the code in the same place. To solve this problem I took the workflow from the datadrivencv package created by Nick Strayer as a basis.

Save your data in Google sheets

We will use the googlesheets4 package to download the cv data from a google spreadsheet. I suggest copying my sheet and replacing its data.

Google Sheets workbook containing structured CV entries The document has the following:

entries: information about education, work experience, certifications, etc.
language_skills: tools or skills you have.
post_info: information about post.
text_block: additional information such as your professional description or interests.
contact_info: contact information (email, phone, etc).

Create custom functions

The following function allows to create an object with all the information of the spreadsheet,

data_location: is the link of the spreadsheet from google sheet
pdf_mode: if TRUE the text links are removed.
sheet_is_publicly_readable: TRUE if the link to the spreadsheet is public.

create_CV_object <-  function(data_location,
                              pdf_mode = FALSE,
                              sheet_is_publicly_readable = TRUE) {
  
  cv <- list(
    pdf_mode = pdf_mode,
    links = c()
  )
  
  is_google_sheets_location <- stringr::str_detect(data_location, "docs\\.google\\.com")
  
  if(is_google_sheets_location){
    if(sheet_is_publicly_readable){
      # This tells google sheets to not try and authenticate. Note that this will only
      # work if your sheet has sharing set to "anyone with link can view"
      googlesheets4::gs4_deauth()
    } else {
      # My info is in a public sheet so there's no need to do authentication but if you want
      # to use a private sheet, then this is the way you need to do it.
      # designate project-specific cache so we can render Rmd without problems
      options(gargle_oauth_cache = ".secrets")
    }
    
    read_gsheet <- function(sheet_id){
      googlesheets4::read_sheet(data_location, sheet = sheet_id, skip = 1, col_types = "c")
    }
    cv$entries_data  <- read_gsheet(sheet_id = "entries")
    cv$skills        <- read_gsheet(sheet_id = "language_skills")
    cv$text_blocks   <- read_gsheet(sheet_id = "text_blocks")
    cv$contact_info  <- read_gsheet(sheet_id = "contact_info")
    cv$post_info  <- read_gsheet(sheet_id = "post_info")
  } else {
    # Want to go old-school with csvs?
    cv$entries_data <- readr::read_csv(paste0(data_location, "entries.csv"), skip = 1)
    cv$skills       <- readr::read_csv(paste0(data_location, "language_skills.csv"), skip = 1)
    cv$text_blocks  <- readr::read_csv(paste0(data_location, "text_blocks.csv"), skip = 1)
    cv$contact_info <- readr::read_csv(paste0(data_location, "contact_info.csv"), skip = 1)
    cv$post_info <- readr::read_csv(paste0(data_location, "post_info.csv"), skip = 1)
  }
  
  
  extract_year <- function(dates){
    date_year <- stringr::str_extract(dates, "(20|19)[0-9]{2}")
    date_year[is.na(date_year)] <- lubridate::year(lubridate::ymd(Sys.Date())) + 10
    
    date_year
  }
  
  parse_dates <- function(dates){
    
    date_month <- stringr::str_extract(dates, "(\\w+|\\d+)(?=(\\s|\\/|-)(20|19)[0-9]{2})")
    date_month[is.na(date_month)] <- "1"
    
    paste("1", date_month, extract_year(dates), sep = "-") %>%
      lubridate::dmy()
  }
  
  # Clean up entries dataframe to format we need it for printing
  cv$entries_data %<>%
    tidyr::unite(
      tidyr::starts_with('description'),
      col = "description_bullets",
      sep = "\n- ",
      na.rm = TRUE
    ) %>%
    dplyr::mutate(description_bullets = as.list(strsplit(description_bullets , "\n- ")) ) %>% 
    dplyr::mutate(
      # description_bullets = ifelse(description_bullets != "", paste0("- ", description_bullets), ""),
      start = ifelse(start == "NULL", NA, start),
      end = ifelse(end == "NULL", NA, end),
      start_year = extract_year(start),
      end_year = extract_year(end),
      no_start = is.na(start),
      has_start = !no_start,
      no_end = is.na(end),
      has_end = !no_end,
      timeline = dplyr::case_when(
        no_start  & no_end  ~ "N/A",
        no_start  & has_end ~ as.character(end),
        has_start & no_end  ~ paste("Current", "-", start),
        TRUE                ~ paste(end, "-", start)
      )
    ) %>%
    dplyr::arrange(desc(parse_dates(end))) %>%
    dplyr::mutate_all(~ ifelse(is.na(.), 'N/A', .))
  
  cv
}

This function removes links from a text

sanitize_links <- function(cv, text){
  if(cv$pdf_mode){
    link_titles <- stringr::str_extract_all(text, '(?<=\\[).+?(?=\\])')[[1]]
    link_destinations <- stringr::str_extract_all(text, '(?<=\\().+?(?=\\))')[[1]]
    
    n_links <- length(cv$links)
    n_new_links <- length(link_titles)
    
    if(n_new_links > 0){
      # add links to links array
      cv$links <- c(cv$links, link_destinations)
      
      # Build map of link destination to superscript
      link_superscript_mappings <- purrr::set_names(
        paste0("<sup>", (1:n_new_links) + n_links, "</sup>"),
        paste0("(", link_destinations, ")")
      )
      
      # Replace the link destination and remove square brackets for title
      text <- text %>%
        stringr::str_replace_all(stringr::fixed(link_superscript_mappings)) %>%
        stringr::str_replace_all('\\[(.+?)\\]', "\\1")
    }
  }
  
  list(cv = cv, text = text)
}

You also need functions to print the downloaded data.

# print text block
print_text_block <- function(cv, label){
  text_block <- dplyr::filter(cv$text_blocks, loc == label) %>%
    dplyr::pull(text)
  
  strip_res <- sanitize_links(cv, text_block)
  
  cat(strip_res$text)
  
  invisible(strip_res$cv)
}

# print table of skills
print_skill <- function(cv){
  
  glue_template <- "
\\cvskill{<<group>>}{<<skills>>}"
  
  skills_formatted <- cv$skills %>%
    mutate(skill = if_else(is.na(details)|details=="", glue::glue("{skill}"), glue::glue("{skill} ({details})"))) %>% 
    group_by(group) %>% 
    summarize(skills = glue::glue_collapse(skill, sep = ", ")) 
  
  cv_skill <- skills_formatted %>% 
    glue::glue_data(glue_template, .open = "<<", .close = ">>" ) %>% 
    paste0(., collapse = '\n')
  
  cv_skill_env <- paste0('\n\\begin{cvskills}\n', cv_skill, '\n\\end{cvskills}\n' , collapse= '\n')
  
  cat(cv_skill_env)
  
  invisible(cv)
}

# print table of posts
print_post<- function(cv, glue_template = "default"){
  
  if(glue_template == "default"){
    glue_template <- "
- [{name}]({link})
\n"
  }  
  print(glue::glue_data(cv$post_info , glue_template))
  
  invisible(cv)
}

Okay, now let’s replace

knitr::opts_chunk$set(echo = FALSE, warning = FALSE, message = FALSE)
library(vitae)

with this, so we can use the previous functions

knitr::opts_chunk$set(echo = FALSE, warning = FALSE, message = FALSE)
library(vitae)
library(dplyr)
library(magrittr)
source("Utils/data_download_funs.R")

CV <- create_CV_object(
  data_location = "https://docs.google.com/spreadsheets/d/19YWFjpiCqTbrwrU1hpxDNubZKoNi-o0DvOQbcHqsc7c/edit?usp=sharing",  
  pdf_mode = FALSE
)

For the Education section we can replace

library(tibble)
tribble(
  ~ Degree, ~ Year, ~ Institution, ~ Where,
  "Informal studies", "1889-91", "Flying University", "Warsaw, Poland",
  "Master of Physics", "1893", "Sorbonne Université", "Paris, France",
  "Master of Mathematics", "1894", "Sorbonne Université", "Paris, France"
) %>% 
  detailed_entries(what = Degree, when = Year, with = Institution, where = Where)

with this

CV$entries_data %>% 
  filter(section == 'education') %>% 
  detailed_entries(
    what = institution,
    when = glue::glue("{`start`} - {`end`}"),
    with = title,
    where = loc,
    why = description_bullets
  )

To print your skills

CV %>% print_skill()

To print your posts

CV %<>% print_post()

All files are hosted here.

I want to thank Nick Strayer and Mitchell O’Hara-Wild for the wonderful work done on the vitae and datadrivencv package.