Title: Unique Location Extractor
Version: 0.1.0
Description: Extracts coordinates of an event location from text based on dictionaries of landmarks, roads, and areas. Only returns the location of an event of interest and ignores other location references; for example, if determining the location of a road traffic crash from the text "crash near [location 1] heading towards [location 2]", only the coordinates of "location 1" would be returned. Moreover, accounts for differences in spelling between how a user references a location and how a location is captured in location dictionaries.
License: MIT + file LICENSE
Encoding: UTF-8
RoxygenNote: 7.3.1
Imports: dplyr, tidyr, readr, purrr, tidytext, stringr, stringi, ngram, hunspell, stringdist, tm, raster, parallel, sf, quanteda, geodist, spacyr, utils
URL: https://dime-worldbank.github.io/ulex/
NeedsCompilation: no
Packaged: 2024-06-16 16:43:41 UTC; robmarty
Author: Robert Marty ORCID iD [aut, cre]
Maintainer: Robert Marty <rmarty@worldbank.org>
Repository: CRAN
Date/Publication: 2024-06-17 18:20:02 UTC

Augments Landmark Gazetteer

Description

Augments Landmark Gazetteer

Usage

augment_gazetteer(
  landmarks,
  landmarks.name_var = "name",
  landmarks.type_var = "type",
  grams.min_words = 3,
  grams.max_words = 6,
  grams.skip_gram_first_last_word_match = TRUE,
  grams.add_only_if_name_new = FALSE,
  grams.add_only_if_specific = FALSE,
  types_rm = c("route", "road", "toilet", "political", "locality", "neighborhood",
    "area", "section of populated place"),
  types_rm.except_with_type = c("flyover", "round about", "roundabout"),
  types_rm.except_with_name = c("flyover", "round about", "roundabout"),
  parallel.sep_slash = TRUE,
  parallel.rm_begin = c(tm::stopwords("en"), c("near", "at", "the", "towards", "near")),
  parallel.rm_end = c("bar", "shops", "restaurant", "sports bar", "hotel", "bus station"),
  parallel.word_diff = "default",
  parallel.word_diff_iftype = list(list(words = c("stage", "bus stop", "bus station"),
    type = "transit_station")),
  parallel.rm_begin_iftype = NULL,
  parallel.rm_end_iftype = list(list(words = c("stage", "bus stop", "bus station"), type
    = "transit_station")),
  parallel.word_begin_addtype = NULL,
  parallel.word_end_addtype = list(list(words = c("stage", "bus stop", "bus station"),
    type = "stage")),
  parallel.add_only_if_name_new = FALSE,
  parallel.add_only_if_specific = FALSE,
  rm.contains = c("road", "rd"),
  rm.name_begin = c(tm::stopwords("en"), c("near", "at", "the", "towards", "near")),
  rm.name_end = c("highway", "road", "rd", "way", "ave", "avenue", "street", "st"),
  pos_rm.all = c("ADJ", "ADP", "ADV", "AUX", "CCONJ", "INTJ", "NUM", "PRON", "SCONJ",
    "VERB", "X"),
  pos_rm.except_type = list(pos = c("NOUN", "PROPN"), type = c("bus", "restaurant",
    "bank"), name = ""),
  close_thresh_km = 1,
  quiet = TRUE
)

Arguments

landmarks

sf spatial points data.frame of landmarks.

landmarks.name_var

Name of variable indicating name of landmark. (Default: "name").

landmarks.type_var

Name of variable indicating type of landmark. (Default: "type").

grams.min_words

Minimum number of words in name to make n/skip-grams out of name. (Default: 3).

grams.max_words

Maximum number of words in name to make n/skip-grams out of name. Setting a cap helps to reduce spurious landmarks that may come out of really long names. (Default: 6).

grams.skip_gram_first_last_word_match

For skip-grams, should first and last word be the same as the original word? (Default: TRUE).

grams.add_only_if_name_new

When creating new landmarks based on n- and skip-grams, only add an additional landmark if the name of the landmark is new; i.e., the name doesn't already exist in the gazetteer. (Default: FALSE).

grams.add_only_if_specific

When creating new landmarks based on n- and skip-grams, only add an additional landmark if the name of the landmark represents a specific location. A specific location is a location where most landmark entries with the same name are close together (within close_thresh_km kilometers). (Default: FALSE).

types_rm

If landmark has one of these types, remove - unless types_rm.except_with_type or types_rm.except_with_name prevents removing. (Default: c("route", "road", "toilet", "political", "locality", "neighborhood", "area", "section of populated place")).

types_rm.except_with_type

Landmark types to always keep. This parameter only becomes relevant in cases where a landmark has more than one type. If a landmark has both a "types_rm" and a "types_always_keep" landmark, this landmark will be kept. (Default: c("flyover", "round about", "roundabout")).

types_rm.except_with_name

Landmark names to always keep. This parameter only becomes relevant in cases where a landmark is one of "types_rm" Here, we keep the landmark if "names_always_keep" is somewhere in the name. For example, if the landmark is a road but has flyover in the name, we may want to keep the landmark as flyovers are small spatial areas. (Default: c("flyover", "round about", "roundabout")).

parallel.sep_slash

If a landmark contains a slash, create new landmarks before and after the slash. (Default: TRUE).

parallel.rm_begin

If a landmark name begins with one of these words, add a landmark that excludes the word. (Default: c(tm::stopwords("en"), c("near","at","the", "towards", "near"))).

parallel.rm_end

If a landmark name ends with one of these words, add a landmark that excludes the word. (Default: c("bar", "shops", "restaurant","sports bar","hotel", "bus station")).

parallel.word_diff

If the landmark includes one of these words, add a landmark that swaps the word for the other word (e.g., "center" with "centre"). By default, uses a set collection of words. Users can also manually specify different word versions. Input should be a data.frame with the following variables: version_1 (for one spelling of the word) and version_2 (for a second spelling of the word).

parallel.word_diff_iftype

If the landmark includes one of these words, add a landmark that swaps the word for the other word (e.g., "bus stop" with "bus station"). Enter a named list of words, with words = c() and type = c(). (Default: list(list(words = c("stage", "bus stop", "bus station"), type = "transit_station"))).

parallel.rm_begin_iftype

If a landmark name begins with one of these words, add a landmark that excludes the word if the landmark is a certain type. (Default: NULL).

parallel.rm_end_iftype

If a landmark name ends with one of these words, add a landmark that excludes the word if the landmark is a certain type. (Default: list(list(words = c("stage", "bus stop", "bus station"), type = "transit_station"))).

parallel.word_begin_addtype

If the landmark begins with one of these words, add the type. For example, if landmark is "restaurant", this indicates the landmark is a restaurant. Adding the "restaurant" to landmark ensures that the type is reflected. (Default: NULL).

parallel.word_end_addtype

If the landmark ends with one of these words, add the type. For example, if landmark is "X stage", this indicates the landmark is a bus stage. Adding the "stage" to landmark ensures that the type is reflected. (Default: list(list(words = c("stage", "bus stop", "bus station"), type = "stage"))).

parallel.add_only_if_name_new

When creating parallel landmarks using the above parameters, only add an additional landmark if the name of the landmark is new; i.e., the name doesn't already exist in the gazetteer. (Default: FALSE).

parallel.add_only_if_specific

When creating parallel landmarks using the above parameters, only add an additional landmark if the name of the landmark represents a specific location. A specific location is a location where most landmark entries with the same name are close together (within close_thresh_km kilometers). (Default: FALSE).

rm.contains

Remove the landmark if it contains one of these words. Implemented after N/skip-grams and parallel landmarks are added. (Default: c("road", "rd")).

rm.name_begin

Remove the landmark if it begins with one of these words. Implemented after N/skip-grams and parallel landmarks are added. (Default: c(tm::stopwords("en"), c("near","at","the", "towards", "near"))).

rm.name_end

Remove the landmark if it ends with one of these words. Implemented after N/skip-grams and parallel landmarks are added. (Default: c("highway", "road", "rd", "way", "ave", "avenue", "street", "st")).

pos_rm.all

Part-of-speech categories to remove. Part-of-speech determined by Spacy. (Default: c("ADJ", "ADP", "ADV", "AUX", "CCONJ", "INTJ", "NUM", "PRON", "SCONJ", "VERB", "X")).

pos_rm.except_type

When specify part-of-speech categories to remove in pos_rm.all, when to override pos_rm.all and keep the word. Names list with: (1) pos (if the word is also another type of part-of-speech); (2) type (if the word is also a certain type of place); and (3) name (if the word includes certain text). Example: list(pos = c("NOUN", "PROPN"), type = c("bus", "restaurant", "bank"), name = c("parliament")). (Default: list(pos = c("NOUN", "PROPN"), type = c("bus", "restaurant", "bank"), name = "")).

close_thresh_km

When to consider locations close together. Used when determining if a landmark name with multiple locations are specific (close together) or general (far apart). (Default: 1).

quiet

Print progress of function. (Default: TRUE).

Value

sf spatial point data.frame of landmarks.

Examples


library(ulex)
library(spacyr)
spacy_install()

lm_sf <- data.frame(name = c("white house",
                             "the world bank group",
                             "the george washington university"),
                    lat = c(38.897778,
                            38.89935,
                            38.9007),
                    lon = c(-77.036389,
                            -77.04275,
                            -77.0508),
                    type = c("building", "building", "building")) |>
sf::st_as_sf(coords = c("lon", "lat"),
         crs = 4326)

lm_aug_sf <- augment_gazetteer(lm_sf)


Locate Event

Description

Locate Event

Usage

locate_event(
  text,
  landmark_gazetteer,
  landmark_gazetteer.name_var = "name",
  landmark_gazetteer.type_var = "type",
  roads,
  roads.name_var = "name",
  areas,
  areas.name_var = "name",
  event_words,
  prepositions_list = list(c("at", "next to", "around", "just after", "opposite", "opp",
    "apa", "hapa", "happened at", "just before", "at the", "outside", "right before"),
    c("near", "after", "toward", "along", "towards", "approach"), c("past", "from",
    "on")),
  junction_words = c("intersection", "junction"),
  false_positive_phrases = "",
  type_list = NULL,
  clost_dist_thresh = 500,
  fuzzy_match = TRUE,
  fuzzy_match.min_word_length = c(5, 11),
  fuzzy_match.dist = c(1, 2),
  fuzzy_match.ngram_max = 3,
  fuzzy_match.first_letters_same = TRUE,
  fuzzy_match.last_letters_same = TRUE,
  quiet = TRUE,
  mc_cores = 1
)

Arguments

text

Vector of texts to be geolocated.

landmark_gazetteer

sf spatial data.frame representing landmarks.

landmark_gazetteer.name_var

Name of variable indicating name of landmark.

landmark_gazetteer.type_var

Name of variable indicating type of landmark.

roads

sf spatial data.frame representing roads.

roads.name_var

Name of variable indicating name of road.

areas

sf spatial data.frame representing areas, such as administrative areas or neighborhoods.

areas.name_var

Name of variable indicating name of area.

event_words

Vector of event words, representing events to be geocoded.

prepositions_list

List of vectors of prepositions. Order of list determines order of preposition precedence. (Default: list(c("at", "next to","around", "just after", "opposite","opp", "apa", "hapa","happened at", "just before","at the","outside", "right before"), c("near", "after", "toward", "along", "towards", "approach"), c("past","from","on"))).

junction_words

Vector of junction words to check for when determining intersection of roads. (Default: c("intersection", "junction")).

false_positive_phrases

Common words found in text that include spurious location references (eg, githurai bus is the name of a bus, but githurai is also a place). These may be common phrases that should be checked and ignored in the text. (Default: "").

type_list

List of vectors of types. Order of list determines order or type precedence. (Default: NULL).

clost_dist_thresh

Distance (meters) as to what is considered "close"; for example, when considering whether a landmark is close to a road. (Default: 500).

fuzzy_match

Whether to implement fuzzy matching of landmarks using levenstein distance. (Default: TRUE).

fuzzy_match.min_word_length

Minimum word length to use for fuzzy matching; vector length must be the same as fuzzy_match.dist. (Default: c(5,11)).

fuzzy_match.dist

Allowable levenstein distances for fuzzy matching; vector length must be same as fuzzy_match.min_word_length. (Default: c(1,2)).

fuzzy_match.ngram_max

The number of n-grams that should be extracted from text to calculate a levensteing distance against landmarks. For example, if the text is composed of 5 words: w1 w2 w3 w4 and fuzzy_match.ngram_max = 3, the function extracts ⁠w1 w2 w3⁠ and compares the levenstein distance to all landmarks. Then in checks ⁠w2 w3 w4⁠, etc. (Default: 3).

fuzzy_match.first_letters_same

When implementing a fuzzy match, should the first letter of the original and found word be the same? (Default: TRUE).

fuzzy_match.last_letters_same

When implementing a fuzzy match, should the last letter of the original and found word be the same? (Default: TRUE).

quiet

If FALSE, prints text that is being geocoded. (Default: TRUE).

mc_cores

If > 1, uses geolocates events in parallel across multiple cores relying on the parallel package. (Default: 1).

Value

sf spatial dataframe of geolocated events.

Examples

library(ulex)
library(sf)

## Landmarks
landmarks_sf <- data.frame(lat = runif(3),
                           lon = runif(3),
                           name = c("restaurant", "bank", "hotel"),
                           type = c("poi", "poi", "poi")) |>
  st_as_sf(coords = c("lon", "lat"),
           crs = 4326)

## Road
coords <- matrix(runif(4), ncol = 2)
road_sf <- coords |>
  st_linestring() |>
  st_sfc(crs = 4326)
road_sf <- st_sf(geometry = road_sf)
road_sf$name <- "main st"

## Area
n <- 5
coords <- matrix(runif(2 * n, min = 0, max = 10), ncol = 2)
coords <- rbind(coords, coords[1,])
polygon <- st_polygon(list(coords))
area_sf <- st_sfc(polygon, crs = 4326)
area_sf <- st_sf(geometry = area_sf)
area_sf$name <- "place"

## Locate Event
event_sf <- locate_event(text = "accident near hotel",
                         landmark_gazetteer = landmarks_sf,
                         roads = road_sf,
                         areas = area_sf,
                         event_words = c("accident", "crash"))