R/functions.R

In CMUSustainability/SDGmapR: Sustainable Development Goal (SDG) Mapping Tools

library(tidyverse)

################ Count functions ################

#' @title Count SDG keywords
#'
#' @description For a given SDG, determine the total weights of the keywords
#' that exist in the text.
#'
#' @param text The target text to map SDG keywords
#' @param sdg The SDG goal we intend to check keywords for
#' @param keywords The specific data set from which to draw keywords (ex. "elsevier", "sdsn")
#' @param count_repeats Whether or not to count repeat weights for the keywords
#'
#' @return A number representing the total weight of words for an SDG
#'
#' @examples
#' count_goal(text = "Poverty entails more than the lack of income and productive resources
#' to ensure sustainable livelihoods.", sdg = 1)
count_sdg_weights <- Vectorize(function(text, sdg, keywords="elsevier100",
                       count_repeats=FALSE) {
    # Initalize total weight
    tot_weight <- 0

    # Select the right keyword set
    if (keywords == "elsevier100") {
      goal_df <- elsevier100_keywords %>%
        filter(goal == sdg)
    } else if (keywords == "sdsn") {
      goal_df <- sdsn_keywords %>%
        filter(goal == sdg)
    } else if (keywords == "elsevier") {
      goal_df <- elsevier_keywords %>%
        filter(goal == sdg)
    } else if (keywords == "cmu250") {
      goal_df <- cmu250_keywords %>%
        filter(goal == sdg)
    } else if (keywords == "cmu500") {
      goal_df <- cmu500_keywords %>%
        filter(goal == sdg)
    } else if (keywords == "cmu1000") {
      goal_df <- cmu1000_keywords %>%
        filter(goal == sdg)
    } else {
      goal_df <- keywords %>%
        filter(goal == sdg)
    }

    if (nrow(goal_df) == 0) {
      return(c())
    }

    # Get the keywords and weights
    goal_patterns <- goal_df$pattern
    goal_weights <- goal_df$weight

    # Add up the keyword weights
    for (idx in 1:nrow(goal_df)) {
      if (str_detect(str_to_lower(text), goal_patterns[idx])) {
        if (count_repeats) {
          keyword_cnt <- str_count(str_to_lower(text), goal_patterns[idx])
          tot_weight <- tot_weight + keyword_cnt * goal_weights[idx]
        } else {
          tot_weight <- tot_weight + goal_weights[idx]
        }
      }
    }

    return (as.numeric(tot_weight))
})

#' @title Count SDGs keywords
#'
#' @description Determine the total weights of all SDGs keywords that exist in a given text
#' for a given goal.
#'
#' @param text The target text to check SDGs keywords
#' @param keywords The specific data set from which to draw keywords (ex. "elsevier", "sdsn")
#' @param count_repeats Whether or not to count repeats for the keywords
#'
#' @return A number representing the weight of words for all SDGs
#'
#' @examples
#' count_sdgs_keywords("Poverty entails more than the lack of income and productive resources
#' to ensure sustainable livelihoods.")
count_sdgs_weights <- function(text, keywords="elsevier2",
                       count_repeats=FALSE) {
  tot_weight <- 0
  for (goal in 1:17) {
    tot_weight <- tot_weight + count_sdg_weights(str_to_lower(text),
                                                 goal,
                                                 keywords, count_repeats)
  }
  return (as.numeric(tot_weight))
}

################ Tabulate functions ################

#' @title Tabulate SDG keywords
#'
#' @description Tabulate all SDG keywords that exist in a given text
#' for a given goal.
#'
#' @param text The target text to check SDG keywords
#' @param sdg The SDG goal we intend to check keywords for
#' @param keywords The specific data set from which to draw keywords (ex. "elsevier", "sdsn")
#' @param count_repeats Whether or not to count repeats for the keywords
#'
#' @return A vector of keywords representing a particular SDG
#'
#' @examples
#' tabulate_sdg_keywords("Poverty entails more than the lack of income and productive resources
#' to ensure sustainable livelihoods.", 1)
tabulate_sdg_keywords <- Vectorize(function(text, sdg, keywords="elsevier100",
                                         count_repeats=FALSE) {

  # Select the right keyword set
  if (keywords == "elsevier100") {
    goal_df <- elsevier100_keywords %>%
      filter(goal == sdg)
  } else if (keywords == "sdsn") {
    goal_df <- sdsn_keywords %>%
      filter(goal == sdg)
  } else if (keywords == "elsevier") {
    goal_df <- elsevier_keywords %>%
      filter(goal == sdg)
  } else if (keywords == "cmu250") {
    goal_df <- cmu250_keywords %>%
      filter(goal == sdg)
  } else if (keywords == "cmu500") {
    goal_df <- cmu500_keywords %>%
      filter(goal == sdg)
  } else if (keywords == "cmu1000") {
    goal_df <- cmu1000_keywords %>%
      filter(goal == sdg)
  } else {
    goal_df <- keywords %>%
      filter(goal == sdg)
  }

  if (nrow(goal_df) == 0) {
    return(c())
  }

  # Get the keywords and weights
  goal_patterns <- goal_df$pattern
  goal_keywords <- goal_df$keyword
  goal_weights <- goal_df$weight

  words <- c()
  # Get keywords in a vector
  for (idx in 1:nrow(goal_df)) {
    if ((str_detect(str_to_lower(text), goal_patterns[idx]))) {
      words <- c(words, goal_keywords[idx])
    }
  }

  return(words)
})

#' @title Tabulate SDG keyword weights
#'
#' @description Tabulate all SDG keyword weights that exist in a given text
#' for a given goal.
#'
#' @param text The target text to check SDG keywords
#' @param goal_num The SDG goal we intend to check keywords for
#' @param keywords The specific data set from which to draw keywords (ex. "elsevier", "sdsn")
#' @param count_repeats Whether or not to count repeats for the keywords
#'
#' @return A vector of weights for keywords representing a particular SDG
#'
#' @examples
#' tabulate_sdg_weights("Poverty entails more than the lack of income and productive resources
#' to ensure sustainable livelihoods.", 1)
tabulate_sdg_weights <- Vectorize(function(text, sdg, keywords="elsevier100",
                                            count_repeats=FALSE) {

  # Select the right keyword set
  if (keywords == "elsevier100") {
    goal_df <- elsevier100_keywords %>%
      filter(goal == sdg)
  } else if (keywords == "sdsn") {
    goal_df <- sdsn_keywords %>%
      filter(goal == sdg)
  } else if (keywords == "elsevier") {
    goal_df <- elsevier_keywords %>%
      filter(goal == sdg)
  } else if (keywords == "cmu250") {
    goal_df <- cmu250_keywords %>%
      filter(goal == sdg)
  } else if (keywords == "cmu500") {
    goal_df <- cmu500_keywords %>%
      filter(goal == sdg)
  } else if (keywords == "cmu1000") {
    goal_df <- cmu1000_keywords %>%
      filter(goal == sdg)
  } else {
    goal_df <- keywords %>%
      filter(goal == sdg)
  }

  if (nrow(goal_df) == 0) {
    return(c())
  }

  # Get the keywords and weights
  goal_patterns <- goal_df$pattern
  goal_keywords <- goal_df$keyword
  goal_weights <- goal_df$weight

  weights <- c()
  # Get keywords in a vector
  for (idx in 1:nrow(goal_df)) {
    if (str_detect(str_to_lower(text), goal_patterns[idx])) {
      weights <- c(weights, goal_weights[idx])
    }
  }

  return (weights)
})

################ Mapping functions ################

#' @title Map text to all SDGs
#'
#' @description Calculate the total weight for all SDGs from 1 to 17
#'
#' @param df Data frame with a column `text`
#' @param keywords The specific data set from which to draw keywords (ex. "elsevier", "sdsn")
#' @param count_repeats Whether or not to count repeats for the keywords
#'
#' @return Original data frame plus 17 additional columns representing weights for
#' SDGs 1 to 17.
#'
#' @examples
#' map_sdgs_weights(df, keywords = "elsevier", count_repeats = TRUE)
map_sdgs_weights <- function(df, keywords="elsevier100", count_repeats=FALSE) {
  new_df <- df %>%
    mutate(`SDG1` = count_sdg_weights(text, 1, keywords, count_repeats),
           `SDG2` = count_sdg_weights(text, 2, keywords, count_repeats),
           `SDG3` = count_sdg_weights(text, 3, keywords, count_repeats),
           `SDG4` = count_sdg_weights(text, 4, keywords, count_repeats),
           `SDG5` = count_sdg_weights(text, 5, keywords, count_repeats),
           `SDG6` = count_sdg_weights(text, 6, keywords, count_repeats),
           `SDG7` = count_sdg_weights(text, 7, keywords, count_repeats),
           `SDG8` = count_sdg_weights(text, 8, keywords, count_repeats),
           `SDG9` = count_sdg_weights(text, 9, keywords, count_repeats),
           `SDG10` = count_sdg_weights(text, 10, keywords, count_repeats),
           `SDG11` = count_sdg_weights(text, 11, keywords, count_repeats),
           `SDG12` = count_sdg_weights(text, 12, keywords, count_repeats),
           `SDG13` = count_sdg_weights(text, 13, keywords, count_repeats),
           `SDG14` = count_sdg_weights(text, 14, keywords, count_repeats),
           `SDG15` = count_sdg_weights(text, 15, keywords, count_repeats),
           `SDG16` = count_sdg_weights(text, 16, keywords, count_repeats),
           `SDG17` = count_sdg_weights(text, 17, keywords, count_repeats))
  return (new_df)
}

#' @title Map text to SDGs keywords
#'
#' @description Return vector of keywords for SDGs 1 to 17
#'
#' @param df Dataframe with a column `text`
#' @param keywords The specific data set from which to draw keywords (ex. "elsevier", "sdsn")
#' @param count_repeats Whether or not to count repeats for the keywords
#'
#' @return Original Dataframe plus 17 additional columns representing vectors of
#' keywords for SDGs 1 to 17.
#'
#' @examples
#' map_sdgs_keywords(df, keywords = "elsevier", count_repeats = TRUE)
map_sdgs_keywords <- function(df, keywords="elsevier100", count_repeats=FALSE) {
  new_df <- df %>%
    mutate(`SDG1_keywords` = tabulate_sdg_keywords(text, 1, keywords, count_repeats),
           `SDG2_keywords` = tabulate_sdg_keywords(text, 2, keywords, count_repeats),
           `SDG3_keywords` = tabulate_sdg_keywords(text, 3, keywords, count_repeats),
           `SDG4_keywords` = tabulate_sdg_keywords(text, 4, keywords, count_repeats),
           `SDG5_keywords` = tabulate_sdg_keywords(text, 5, keywords, count_repeats),
           `SDG6_keywords` = tabulate_sdg_keywords(text, 6, keywords, count_repeats),
           `SDG7_keywords` = tabulate_sdg_keywords(text, 7, keywords, count_repeats),
           `SDG8_keywords` = tabulate_sdg_keywords(text, 8, keywords, count_repeats),
           `SDG9_keywords` = tabulate_sdg_keywords(text, 9, keywords, count_repeats),
           `SDG10_keywords` = tabulate_sdg_keywords(text, 10, keywords, count_repeats),
           `SDG11_keywords` = tabulate_sdg_keywords(text, 11, keywords, count_repeats),
           `SDG12_keywords` = tabulate_sdg_keywords(text, 12, keywords, count_repeats),
           `SDG13_keywords` = tabulate_sdg_keywords(text, 13, keywords, count_repeats),
           `SDG14_keywords` = tabulate_sdg_keywords(text, 14, keywords, count_repeats),
           `SDG15_keywords` = tabulate_sdg_keywords(text, 15, keywords, count_repeats),
           `SDG16_keywords` = tabulate_sdg_keywords(text, 16, keywords, count_repeats),
           `SDG17_keywords` = tabulate_sdg_keywords(text, 17, keywords, count_repeats))
  return (new_df)
}