plot_spatial_betas {BKTR}

Plot Spatial Beta Coefficients

Description

Create a plot of beta values through space for a given temporal point and a set of feature labels.

Usage

plot_spatial_betas(
  bktr_reg,
  plot_feature_labels,
  temporal_point_label,
  nb_cols = 1,
  use_dark_mode = TRUE,
  show_figure = TRUE,
  zoom = 11,
  google_token = NULL,
  fig_width = 8.5,
  fig_height = 5.5,
  fig_resolution = 200
)

Arguments

bktr_reg

BKTRRegressor: BKTRRegressor object.

plot_feature_labels

Array: Array of feature labels to plot.

temporal_point_label

String: Temporal point label to plot.

nb_cols

Integer: The number of columns to use in the facet grid.

use_dark_mode

Boolean: Whether to use a dark mode for the geographic map or not.

show_figure

Boolean: Whether to show the figure. Defaults to True.

zoom

Integer: Zoom level for the geographic map. Defaults to 11.

google_token

String or NULL: Google API token to use for the geographic map. Defaults to NULL. If NULL, use Stamen maps.

fig_width

Numeric: Figure width when figure is shown. Defaults to 8.5.

fig_height

Numeric: Figure height when figure is shown. Defaults to 5.5.

fig_resolution

Numeric: Figure resolution PPI. Defaults to 200.

Value

ggplot or NULL: ggplot object or NULL if show_figure is set to FALSE.

Examples

# Launch MCMC sampling on a light version of the BIXI dataset
bixi_data <- BixiData$new(is_light = TRUE)
bktr_regressor <- BKTRRegressor$new(
  data_df <- bixi_data$data_df,
  spatial_positions_df = bixi_data$spatial_positions_df,
  temporal_positions_df = bixi_data$temporal_positions_df,
  burn_in_iter = 5, sampling_iter = 10) # For example only (too few iterations)
bktr_regressor$mcmc_sampling()

# Plot spatial beta coefficients for the first time point and the two features
plot_spatial_betas(
  bktr_regressor,
  plot_feature_labels = c('mean_temp_c', 'area_park'),
  temporal_point_label = bixi_data$temporal_positions_df$time[1])

# We can also use light mode and plot the maps side by side
plot_spatial_betas(
  bktr_regressor,
  plot_feature_labels = c('mean_temp_c', 'area_park', 'total_precip_mm'),
  temporal_point_label = bixi_data$temporal_positions_df$time[10],
  use_dark_mode = FALSE,  nb_cols = 3)