Skip to contents

Converting point and event data

Source: vignettes/pg_from_points.Rmd
pg_from_points.Rmd

pg_from_points() converts coordinate tables, sf objects, SpatVector objects, and point-vector files to cell-level data. Use it for event counts, presence indicators, or summaries of a numeric value.

Example data and grid 

events <- data.frame(
  longitude = c(10.1, 10.4, 20.2, 20.4),
  latitude = c(50.1, 50.4, 60.2, 60.4),
  event_date = as.Date(c(
    "2020-03-01", "2020-07-15", "2021-02-01", "2021-09-10"
  )),
  fatalities = c(2, 3, 1, 4),
  event_type = c("A", "A", "B", "B")
)

config <- priogrid::pg_config(
  nrow = 18,
  ncol = 36,
  temporal_resolution = "1 year",
  start_date = as.Date("2020-12-31"),
  end_date = as.Date("2021-12-31"),
  verbose = FALSE
)

Count events

Specify coordinate columns when x is a data frame. Coordinates are assumed to use EPSG:4326 unless input_crs says otherwise.

event_counts <- pg_from_points(
  x = events,
  variable = "event_count",
  longitude = "longitude",
  latitude = "latitude",
  aggregation = "count",
  time_column = "event_date",
  config = config
)

event_counts
#>   pgid measurement_date event_count
#> 1  524       2020-12-31           2
#> 2  561       2021-12-31           2

Aggregate values

The sum, mean, min, and max rules require a value column.

fatality_totals <- pg_from_points(
  x = events,
  variable = "fatalities",
  longitude = "longitude",
  latitude = "latitude",
  value = "fatalities",
  aggregation = "sum",
  time_column = "event_date",
  config = config
)

fatality_totals
#>   pgid measurement_date fatalities
#> 1  524       2020-12-31          5
#> 2  561       2021-12-31          5

Groups and empty cells

Use group_columns to retain separate summaries, for example by event type. Grouped results support return = "data" only.

counts_by_type <- pg_from_points(
  x = events,
  variable = "event_count",
  longitude = "longitude",
  latitude = "latitude",
  aggregation = "count",
  group_columns = "event_type",
  config = config
)

With keep_empty = TRUE, unobserved cells receive zero only for count, presence, and sum. Missing means, minima, and maxima remain missing because zero would have a different substantive meaning.

Spatial inputs and output

If x is already an sf or SpatVector point object, omit longitude, latitude, and input_crs. Set return = "result" when both tabular and raster representations plus metadata are useful.

result <- pg_from_points(
  events,
  variable = "event_count",
  longitude = "longitude",
  latitude = "latitude",
  aggregation = "count",
  return = "result",
  config = config
)

pg_write_result(result, "event_counts.csv")

See vignette("pg_from_vector", package = "priogridR") for mixed vector workflows involving lines or polygons.