predict_locs: predict_locs

In raffscallion/rapidfire: relatively accurate particulate information derived from inputs retrieved easily

predict_locs

Description

Uses an existing model, such as developed by develop_model, to predict values a given locations and dates. Prepares the data necessary to do so. MAIAC, NARR, BlueSky, and (optionally) PurpleAir data should already be downloaded.

Usage

predict_locs(
  dt1,
  dt2,
  states = "CA",
  model,
  locations,
  pa_cutoff = 1e+05,
  bluesky_special = NULL,
  pa_data = NULL
)

Arguments

dt1	Date The earliest date of data to predict
dt2	Date The latest date to predict
states	character A vector of two-character state codes
model	A final model object, such as extracted from the results of develop_model. For example, if the object was called mod, this would be mod$model$finalModel
locations	A SpatialPointsDataFrame of locations (and dates) to predict
pa_cutoff	A cutoff value passed to vgm for interploation of PurpleAir data
bluesky_special	character Handles three special cases of BlueSky data. If "2020", processes data prior to October 10, 2020 separately from that after, as the BlueSky data format changed. If "HAQAST", uses custom BlueSky-CMAQ output created during the HAQAST campaign (see https://doi.org/10.1080/10962247.2021.1891994O'Neill et al., 2021). If "nominal", instead use a nominal placeholder value for BlueSky PM2.5 of 0.1, which will have a near neutral impact on predictions.
pa_data	PurpleAir data in pas-like format.

Value

A dataframe with all model input values and the resulting predictions for the locations and dates specified in locations

Examples

pred <- predict_locs(dt1, dt2, states = "CA", model = mod$model$finalModel,
                               locations = locs, pa_data = pa)

raffscallion/rapidfire documentation built on June 1, 2025, 2:04 p.m.