In hrvg/regionaldrivers: Statistical analysis of regional drivers for classifications

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.width = 10,
  fig.height = 8,
  message = FALSE,
  warning = FALSE
)

Libraries

library(regionaldrivers)
library(magrittr)

Data loading

SSCT_data and SSCT_labelled_data are loaded with the package. Here is what 10 random rows of each looks like:

SSCT_labelled_data %>% dplyr::sample_n(10) %>%
    knitr::kable(digits = 3, format = "html", caption = "SSCT_labelled_data") %>% 
    kableExtra::kable_styling(bootstrap_options = c("hover", "condensed")) %>% 
    kableExtra::scroll_box(width = "7in", height = "5in")

SSCT_data %>% dplyr::sample_n(10) %>%
    knitr::kable(digits = 3, format = "html", caption = "SSCT_data") %>% 
    kableExtra::kable_styling(bootstrap_options = c("hover", "condensed")) %>% 
    kableExtra::scroll_box(width = "7in", height = "5in")   

Data transformation

We transform the data to retain what we need for the analysis:

drivers_data <- cbind(SSCT_labelled_data, SSCT_data) %>%
    dplyr::select(c("layer_sd.rstr", "slope_mean.rstr", "SLOPE", "RUSLE", "ward"))  %>%
    # dplyr::select(c("CONFINEMEN", "layer_sd.rstr", "H.960", "slope_mean.rstr", "SLOPE", "RUSLE", "ward")) %>%
    dplyr::rename(ward.grp = ward) %>%
    dplyr::mutate(RUSLE = log10(RUSLE)) %>%
    reshape2::melt(id.vars = "ward.grp")
head(drivers_data)

Statistical testing and visualization

We can now visualize the results from a pair-wise statistical comparison with significative_diff_boxplot().

Box plots

Dunn_results <- significative_diff_boxplot(drivers_data, type = "Dunn", padjmeth = "bonferroni", pvalue = 0.10, no_label = TRUE)
Dunn_results$p

Matrices

As the boxplots might be difficult to read, the p-value can be recasted into a matrix form with plot_pairwise_pvalue().

plotly::ggplotly(
    plot_pairwise_pvalue(Dunn_results$stats),
    tooltip = c("group1", "group2", "p.value")
)   

Lollipop charts

We can repeat the above procedure for all data and visualize the results at the number of significant pairwise comparisons. This time we select most data, removing the data with near-zero variance and the Indices of Catchment Integrity. significative_count_lollipop() produces a lollipop chart counting the number of significant pairwise comparison. If normalize = TRUE, the counts are normalized by the number of pairwise comparisons.

all_drivers_data <- cbind(SSCT_labelled_data, SSCT_data %>% dplyr::select("ward")) %>%
    dplyr::select(-caret::nearZeroVar(., names = TRUE)) %>%
    dplyr::select(-c("CHYD", "CCHEM", "CSED", "CCONN", "CTEMP", "CHABT", "ICI", "WHYD", "WCHEM", "WSED", "WCONN", "WTEMP", "WHABT")) %>%
    dplyr::select(-dplyr::contains("H.")) %>%
    dplyr::rename(ward.grp = ward) %>%
    dplyr::mutate(RUSLE = log10(RUSLE)) %>%
    reshape2::melt(id.vars = "ward.grp") %>%
    na.omit()
Dunn_results <- significative_diff_boxplot(all_drivers_data, type = "Dunn", padjmeth = "bonferroni", pvalue = 0.10)
significative_count_lollipop(Dunn_results, pvalue = 0.10, first = 15)$p %>% plotly::ggplotly() 
significative_count_lollipop(Dunn_results, pvalue = 0.10, normalize = TRUE, first = 15)$p %>% plotly::ggplotly()

hrvg/regionaldrivers documentation built on June 20, 2021, 7:50 a.m.