Bedny et al. (2019) Human- Language comparision

library(tidyverse)
library(here)
library(knitr)
library(broom)
library(Matrix)
library(dendextend)
library(cowplot)


TAXONOMIC_PATH <- here("data/raw/taxonomy_matrix.mat")
taxonomic_data <- R.matlab::readMat(TAXONOMIC_PATH)[[2]]  
LABELS <- c("shark", "swan", "flamingo", "pigeon", "crow", "elephant", 
            "mammoth", "sloth", "beaver", "gorilla", "bat", "rhino", 
            "zebra", "llama", "hippo", "killerwhale", "dolphin", "giraffe",
            "sheep", "goat", "deer", "pig", "boar", "lion", "panther", "cheetah",
            "skunk", "panda", "polarbear", "grizzly") %>% rev() # from SI fig s2
colnames(taxonomic_data) <- LABELS
rownames(taxonomic_data) <- LABELS
taxonomic_long <- taxonomic_data %>%
  as.data.frame() %>%
  rownames_to_column("animal1") %>%
  gather("animal2", "similarity", -animal1) %>%
  mutate(sim_type = "taxonomic_similarity")

LANGUAGE_PATH <- here("data/processed/animal_distances_coca_all_cleaned_gary.csv")

language_data <- read_csv(LANGUAGE_PATH) %>%
  spread(word2, language_similarity) %>%
  select(-word1)

all_corrs_mat_langs <- as.matrix(language_data)
rownames(all_corrs_mat_langs) <- colnames(all_corrs_mat_langs)
language_long <- all_corrs_mat_langs %>%
  as.data.frame() %>%
  rownames_to_column("animal1") %>%
  gather("animal2", "similarity", -animal1)   %>%
  mutate(sim_type = "language_coca_all")

taxo_corr <- bind_rows(taxonomic_long, language_long) %>%
  filter(animal1 < animal2) %>%
  spread(sim_type, similarity)  %>%
  mutate(participant_type = "Ground Truth") %>%
  group_by(participant_type) %>%
  nest() %>%
  mutate(temp = map(data, ~ tidy(cor.test(.$language_coca_all,
                                          -.$taxonomic_similarity, method = "spearman"))),
         n = map(data, nrow),
         dimension = "Taxonomy")  %>%
  select(-data) %>%
  unnest()  %>%
    mutate(se = 1/sqrt(n-3),
         estimate_se_l = estimate - se,
         estimate_se_h = estimate + se,)

kable(taxo_corr)

LANGUAGE_PATH <- here("data/processed/animal_distances_coca_all_cleaned_animal_gary.csv")

language_data <- read_csv(LANGUAGE_PATH) %>%
  spread(word2, language_similarity) %>%
  select(-word1)

all_corrs_mat_langs <- as.matrix(language_data)
rownames(all_corrs_mat_langs) <- colnames(all_corrs_mat_langs)
language_long <- all_corrs_mat_langs %>%
  as.data.frame() %>%
  rownames_to_column("animal1") %>%
  gather("animal2", "similarity", -animal1)   %>%
  mutate(sim_type = "language_coca_al_animal")

taxo_corr <- bind_rows(taxonomic_long, language_long) %>%
  filter(animal1 < animal2) %>%
  spread(sim_type, similarity)  %>%
  mutate(participant_type = "Ground Truth") %>%
  group_by(participant_type) %>%
  nest() %>%
  mutate(temp = map(data, ~ tidy(cor.test(.$language_coca_al_animal,
                                          -.$taxonomic_similarity,
                                          method = "spearman"))),
         n = map(data, nrow),
         dimension = "Taxonomy")  %>%
  select(-data) %>%
  unnest()  %>%
  mutate(se = 1/sqrt(n-3),
         estimate_se_l = estimate - se,
         estimate_se_h = estimate + se,)

kable(taxo_corr)